Free LLM API Tokens
Updated daily

Free API credits for AI models, in one place.

This site is a daily-updated tracker of 124 free and free-tier LLM APIs. For each provider it records the published free allowance, whether the API is OpenAI-compatible, the base URL, and tokens per day — or “Not published” when official pages do not give a number.

Last updated 14 Nov 2026.

124providers listed
14 Nov 2026, 03:44 UTClast checked (UTC)

Figures come from each provider’s own pages on the date shown, and they change often. Check before you rely on them. Never share your API keys.

How to get started

Sign up · Get a key · Set the base URL
1

How do I sign up?

Open a provider from the list and create a free account. Prefer official providers if you are sending anything private.

2

Where do I get an API key?

Copy the key from their dashboard and keep it secret — on your computer or in an environment variable, never in a public chat or repo.

3

How do I set the base URL?

Point the OpenAI SDK (or any OpenAI-compatible client) at the provider’s base URL:

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.API_KEY,
  baseURL: "https://api.example.com/v1",
});

Which sites currently offer free LLM API tokens?

More filters
Latest updates

13 Nov 2026

12 Nov 2026

11 Nov 2026

10 Nov 2026

9 Nov 2026

See the full What’s new page or subscribe to the RSS feed.

OpenAI compatibleOfficial provider
What you get free
Complimentary Token Lab points for free AMD AI Developer Program members (10 points, 30-day access per AMD dev community)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.tokenlab.amd.com/v1
Limits
About 20 RPM per key; points drawn per model rate; access extendable once by 15 days
Notes
Open models on AMD GPUs: DeepSeek V4, Qwen3.8, Gemma 4, GLM-5.3. AMD publishes no point amount on the program page.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.tokenlab.amd.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
14 Nov 2026
OpenAI compatibleRelay warning
What you get free
$1 welcome credit on signup (first 100 accounts) plus free-tier Flash model
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://nflashx.cc.cd/v1
Limits
Flash model 2 req/min on free, 4K output
Notes
Small new multimodal API (chat, image, video) by NowayDev; credit capped per network.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://nflashx.cc.cd/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
14 Nov 2026
OpenAI compatibleRelay warning
What you get free
$5 trial credit on request (email support or DM @QoraAPI); not automatic
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://qoraapi.com/v1
Limits
Pay as you go after; $50 minimum top-up
Notes
API relay gateway, 350+ models, also Anthropic-format at https://qoraapi.com.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://qoraapi.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
14 Nov 2026
OpenAI compatibleOfficial providerOngoing free
What you get free
Starter plan is $0/month and includes $10 of model credits each month for built-in models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
ongoing
Base URL
https://api.braintrust.dev/v1/proxy
Limits
Monthly credits do not roll over. On Starter, built-in model use pauses after credits unless a card is added for on-demand usage
Notes
Official examples still use this OpenAI-compatible proxy URL; docs also note the proxy page is deprecated and that requests now route to Braintrust's gateway.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.braintrust.dev/v1/proxy",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
13 Nov 2026
OpenAI compatibleRelay warningTrial / one-time
What you get free
1M free tokens one-time for new accounts (email or Google sign-in)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://api.gonkarouter.io/v1
Limits
Not published
Notes
Router on the Gonka network; GLM-5.3-Flash, DeepSeek-V4-Flash, MiniMax-M2.7. Also Anthropic-compatible. One-time grant, not daily. Blog mentions a $20 credit campaign that the pricing page does not confirm.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.gonkarouter.io/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
13 Nov 2026
OpenAI: not publishedOfficial providerOngoing free
What you get free
Free plan: up to 300,000 foundation-model tokens per month, up to 20 CUH of machine learning tools, and up to 100 documents for text extraction
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
ongoing
Base URL
https://us-south.ml.cloud.ibm.com
Limits
Monthly Free-plan caps above; a tokens-per-day figure is not published
Notes
Dallas service URL is from IBM Cloud API Docs; other regions have their own hosts. A simple OpenAI-compatible base URL was not clearly published on the pages checked.
Source
Official page
·
Last verified
13 Nov 2026
OpenAI compatibleOfficial providerTrial / one-time
What you get free
Limited-time $2 in free API credits for new accounts (stated on console.liquid.ai)
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
trial
Base URL
https://api.liquid.ai/openai/v1
Limits
Per-model request and token limits are listed in Liquid Console, not as a single public number
Notes
Official docs say chat models speak the OpenAI API. The $2 new-account credit is described as limited-time on the console.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.liquid.ai/openai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
13 Nov 2026
Not OpenAI compatibleOfficial providerOngoing free
What you get free
Free plan can test the I models without a credit card. The pay-as-you-go plan adds an initial $15 credit and requires a card
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
ongoing
Base URL
https://api.nlpcloud.io/v1
Limits
Official FAQ says throughput on the Free plan is very limited; no public numeric cap was found
Notes
Uses NLP Cloud's own API, not an OpenAI-compatible chat endpoint. The $15 credit applies only to pay-as-you-go.
Source
Official page
·
Last verified
13 Nov 2026
OpenAI compatibleOfficial providerOngoing free
What you get free
Apodex 1.1 Mini is listed with Free input and Free output on the official pricing page
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
ongoing
Base URL
https://api.novita.ai/openai/v1
Limits
Not published
Notes
Official quickstart uses the OpenAI-compatible chat completions endpoint. Other catalog models on the pricing page are paid.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.novita.ai/openai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
13 Nov 2026
OpenAI compatibleRelay warningOngoing free
What you get free
Free plan includes $1 of model usage each month and access to all open catalog models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
ongoing
Base URL
https://speka.me/v1
Limits
10 requests/minute and 1 API key on the Free plan
Notes
Official site says prompts are not used to train models. A gateway: do not send private data if that is a concern.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://speka.me/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
13 Nov 2026
OpenAI compatibleOfficial providerOngoing free
What you get free
Ternary Bonsai 27B (Prism-ML/Ternary-Bonsai-27B) is listed as Free for input and output on the serverless models page
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
ongoing
Base URL
https://api.together.ai/v1
Limits
No fixed public RPM for this model; official rate-limit docs say limits may apply under high demand
Notes
Other Together serverless models on the same page are priced per token. Confirm the current Free listing before you rely on it.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.together.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
13 Nov 2026
OpenAI compatibleRelay warningTrial / one-time
What you get free
$5 one-time welcome credit after CLI login and project link; then pay-as-you-go wallet
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://vectoree.ai/api/v1
Limits
Not published
Notes
Agent backend with OpenAI-compatible Model Gateway (36+ models, vectoree/auto). Credit is one-time, not daily; empty wallet stops the gateway.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://vectoree.ai/api/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
13 Nov 2026
OpenAI compatibleRelay warningTrial / one-time
What you get free
$2 signup credit after email verify; 5 free Zhipu Flash models (glm-4.7-flash etc.)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://aiapi-pro.com/v1
Limits
10 free-model calls/day, 5 RPM before first top-up
Notes
Gateway relaying Chinese providers; no daily token figure published.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://aiapi-pro.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
12 Nov 2026
OpenAI compatible
What you get free
Free plan: daily usage on DeepSeek V4.1 Flash only (GLM-5.3 Flash no longer free)
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://zro.moonmath.ai/v1
Limits
Not published
Notes
Daily amount not published; zero request retention; EU/US inference.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://zro.moonmath.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
12 Nov 2026
OpenAI compatibleRelay warning
What you get free
$1 free credit on signup, 100+ models on one key (15% markup after)
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.1vy.ai/v1
Limits
Not published
Notes
Unofficial multi-model gateway.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.1vy.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
11 Nov 2026
OpenAI compatibleRelay warning
What you get free
$10 free credit on signup, 53 models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
Not published
Limits
100 req/min on Free; tokens/day unspecified
Notes
Multi-model gateway, OpenAI-compatible format per its page; base URL not stated on homepage. Young unofficial gateway.
Source
Official page
·
Last verified
11 Nov 2026
OpenAI compatibleOfficial providerTrial / one-time
What you get free
One-time $1 credit after phone verification (unlocks inference on Free Tier); credits never expire once bought
Tokens per day
Not published
Card needed
No
Phone needed
Yes
Offer type
trial
Base URL
https://api.hyperbolic.ai/inference/v1
Limits
300 req/min per key (per docs via third-party summaries); tokens/day unspecified
Notes
Relaunched serverless API; old api.hyperbolic.xyz is retired. Text and vision models only. Each phone number works for one account.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.hyperbolic.ai/inference/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
11 Nov 2026
OpenAI compatible
What you get free
Free for research/education, 'generous' quota, tokens/day unpublished
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://freeinference.org/v1
Limits
Not published
Notes
Email signup. OpenAI and Anthropic compatible endpoints. Research lab, so avoid private data.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://freeinference.org/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
10 Nov 2026
OpenAI compatibleRelay warning
What you get free
Free plan: 5,000 requests/month, 100 managed requests and $0.50 model usage/day on models up to $2/M output tokens
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
Not published
Limits
60 RPM per key, 10 RPM managed, 8,192 in / 1,024 out tokens per request
Notes
Gateway with smart routing; tokens/day unspecified (daily cap is in dollars).
Source
Official page
·
Last verified
10 Nov 2026
OpenAI compatibleOfficial provider
What you get free
K-EXAONE 2.5 API free to use with an API key (announced 9 Nov 2026); no quota published
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.exaone.ai/v1
Limits
Not published
Notes
Official LG AI Research. Model k-exaone_v2.5, 512K context, Korean/English text only. Rate limits and end date not published.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.exaone.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
9 Nov 2026
OpenAI compatible1M / dayTrial / one-time
What you get free
30-day free trial, 1M tokens/day on all Core models
Tokens per day
1,000,000
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://api.regolo.ai/v1
Limits
Stricter rate limits, fair-usage throttling
Notes
Italian EU hosting. Trial ends after 30 days, then pay-as-you-go. Model multipliers affect how fast the daily allowance is used.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.regolo.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
9 Nov 2026
OpenAI compatibleRelay warning
What you get free
$5 signup bonus (Starter model pool); may expire after 30 days if unused
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.boundlessapi.com/v1
Limits
Starter pool rate limits; tokens/day unspecified
Notes
Unified router for Claude/GPT/Gemini and curated models. Billing policy also notes a 5% platform fee on credit purchases after the bonus.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.boundlessapi.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
6 Nov 2026
OpenAI compatibleRelay warning
What you get free
$5 promotional inference credit per new organisation; expires in 90 days; one grant per org
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.llmeu.com/v1
Limits
Default policy caps $10/org/day; tokens/day unspecified
Notes
European catalogue/router. Partner endpoints (including OpenRouter rows) may need policy turn-on. Card payments not enabled on this deployment.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.llmeu.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
6 Nov 2026
OpenAI compatibleTrial / one-time
What you get free
$1 free credits on signup; no card required
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://api.llmcloud.ai/v1
Limits
Tokens/day unspecified
Notes
One OpenAI-compatible key across 150+ models with auto:cost / auto:quality / auto:reasoning routing. Signup page states free $1 credit.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.llmcloud.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
6 Nov 2026
OpenAI compatibleOfficial provider
What you get free
$10 free credits on signup (launch offer); 30% off models through 11 Nov 2026
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.mangoboost.io/v1
Limits
Tokens/day unspecified; credits expire after stated validity period
Notes
MangoBoost first-party OpenAI-compatible (and Anthropic Messages) API for open-weight models such as DeepSeek V4.1 Flash, GLM-5.3, Kimi-K3. Korean infrastructure.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.mangoboost.io/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
6 Nov 2026
OpenAI compatibleOfficial provider
What you get free
S5-free model: no token charge for a limited launch window (text only; no Batch/Flex/built-in tools)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://chat.maritaca.ai/api
Limits
S5-free: 60 RPM, 256K TPM in, 256K TPM out (fixed across tiers)
Notes
Brazilian Sabiá/S5 API. Official pricing docs (maritalk-api) document S5-free as temporary. Paid S5 remains available when free ends. Base URL from Maritaca API docs / community examples.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://chat.maritaca.ai/api",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
6 Nov 2026
OpenAI compatibleOfficial provider
What you get free
$5 free starter credit for new accounts (stated in docs changelog for Serverless Inference and Fine-tuning); quickstart says early users may be eligible
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://inference.api.nscale.com/v1
Limits
Tokens/day unspecified; 429 on insufficient credit
Notes
UK AI cloud OpenAI-compatible serverless inference. Google SSO signup. Confirm free credit still appears in console on signup.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://inference.api.nscale.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
6 Nov 2026
OpenAI compatibleRelay warning
What you get free
Free models bill $0 per token (OpenRouter :free routes plus OpenPaths Free and others); paid models need credits
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://openpaths.io/v1
Limits
Best-effort rate limits per key; tokens/day unspecified
Notes
Open-source model router. Free catalogue lists many or/*:free models plus nvidia/deepseek-v4-flash and openpaths/free. No card for free routes.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://openpaths.io/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
6 Nov 2026
OpenAI compatibleOfficial provider
What you get free
Free of charge while the API is experimental; models Qwen3.6-35B-A3B-FP8 and Qwen3.8-27B
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://inference.hetzner.com/api/v1
Limits
Per key per 60s: 10 requests, 4M input tokens, 100K output tokens; no daily token cap published
Notes
Needs a Hetzner account; create a token at experiments.hetzner.com. Not for production; no availability guarantee.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://inference.hetzner.com/api/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
5 Nov 2026
OpenAI compatible
What you get free
5 free credits on signup (25 more if you add a card); docs say up to 40M free tokens to start; credits expire after 12 months
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://lightning.ai/api/v1
Limits
Free: 15 RPM, 120K tokens/min
Notes
One key across GPT, Claude, Gemini, DeepSeek, GLM and more. Personal-email accounts can take 1-3 days to activate.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://lightning.ai/api/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
5 Nov 2026
OpenAI compatibleRelay warning
What you get free
Free key: 100 requests/day plus one-time +400 welcome bonus (per the team's 3 Nov launch post)
Tokens per day
Not published
Card needed
No card
Phone needed
Not published
Offer type
Not published
Base URL
https://api.mervex.cc.cd/v1
Limits
Extra daily quota for linking Discord/Google/GitHub
Notes
Small bootstrapped gateway on a free subdomain, 100+ open models (Gemma, Llama, Qwen, GLM, Kimi, DeepSeek)
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.mervex.cc.cd/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
4 Nov 2026
OpenAI compatibleOfficial provider
What you get free
Free plan: 50 credits (₹50) on signup, then 50 every month; GPT OSS 20B only
Tokens per day
Not published
Card needed
No card needed
Phone needed
Not published
Offer type
Not published
Base URL
https://api.nabh.cloud/v1
Limits
5 requests/min, 500 requests/day, 1 API key
Notes
Indian prepaid LLM marketplace billed in rupees; top-ups via UPI from ₹500
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.nabh.cloud/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
4 Nov 2026
OpenAI compatibleRelay warningTrial / one-time
What you get free
$0.50 signup credit (homepage; launch post said $1), does not expire
Tokens per day
Not published
Card needed
No card for the signup credit
Phone needed
Not published
Offer type
trial
Base URL
https://api.openany.si/v1
Limits
Prepaid pay-per-call after the credit; ~154 models plus data APIs
Notes
Launched October 2026
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.openany.si/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
4 Nov 2026
OpenAI compatibleOfficial provider
What you get free
$1 free credit on signup
Tokens per day
Not published
Card needed
No card
Phone needed
Not published
Offer type
Not published
Base URL
https://my.orq.ai/v3/router
Limits
Pay as you go after the credit; 500+ models
Notes
EU-hosted gateway, one key for OpenAI, Anthropic, Google, Mistral and more
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://my.orq.ai/v3/router",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
4 Nov 2026
OpenAI compatibleRelay warningTrial / one-time
What you get free
3M one-time starter tokens after email confirmation
Tokens per day
Not published
Card needed
No card
Phone needed
Not published
Offer type
trial
Base URL
https://gate.joingonka.ai/v1
Limits
Models: GLM-5.3-Flash, DeepSeek-V4-Flash, MiniMax-M2.7; then pay per token
Notes
Also Anthropic-format at https://gate.joingonka.ai. One-time bonus, not a daily quota.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://gate.joingonka.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Nov 2026
OpenAI compatibleOfficial providerTrial / one-time
What you get free
$1 credit for new accounts
Tokens per day
Not published
Card needed
No card stated for the $1 credit
Phone needed
Not published
Offer type
trial
Base URL
https://gateway.ngrok.ai/v1
Limits
ngrok-managed keys cover OpenAI and Anthropic models; credits otherwise prepaid (min $5)
Notes
Sign in at app.ngrok.ai. Bring-your-own-key also supported.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://gateway.ngrok.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Nov 2026
OpenAI compatibleRelay warningTrial / one-time
What you get free
$10 signup credit (one-time) on any listed model
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://helyxai.space/v1
Limits
Not published
Notes
Third-party aggregator listings say $2 is instant and $8 is released by support; daily free-token claims are mentioned but no figure is published on the official site. Crypto top-ups.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://helyxai.space/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
2 Nov 2026
OpenAI compatibleOfficial provider
What you get free
€5 free API credit on signup (expires after 90 days)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.infercom.ai/v1
Limits
Standard free-plan rate limits (see docs.infercom.ai rate limits)
Notes
EU-hosted SambaNova Cloud service; production models only on Free.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.infercom.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
2 Nov 2026
OpenAI: not publishedOfficial providerTrial / one-time
What you get free
1M free tokens per language model, one-time, valid 1 year (campaign until 31 Dec 2026)
Tokens per day
Not published
Card needed
Top-up
Phone needed
Not published
Offer type
trial
Base URL
Not published
Limits
Not published
Notes
Each model claim freezes $10 of account balance or credit limit, returned when endpoints are deleted.
Source
Official page
·
Last verified
2 Nov 2026
OpenAI compatibleRelay warning
What you get free
$100 free credits on signup via GitHub; 3 Free Tier models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://aevon.sh/v1
Limits
Unpublished
Notes
Gateway on Venice AI infra, has its own crypto token; also Anthropic base URL. Treat as prototype-only.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://aevon.sh/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
30 Oct 2026
OpenAI compatibleRelay warning
What you get free
Free plan $0/month on basic models (e.g. gpt-oss-20b)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.airforce/v1
Limits
1 request/min, 1,000 requests/day
Notes
Multi-model gateway, 100+ models on paid plans.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.airforce/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
30 Oct 2026
OpenAI compatibleRelay warning
What you get free
$1 on email verify, then one daily claim (up to $20 over first 40 days, continues after)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://aiapiv2.pekpik.com/v1
Limits
10 requests/min on free tier; daily claim resets balance, no stacking
Notes
Multi-model gateway; one permanent key.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://aiapiv2.pekpik.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
30 Oct 2026
OpenAI compatibleOfficial provider
What you get free
Free public-beta Token Plan: shared credit pool across Kimi K3, GLM-5.2, DeepSeek V4.1 Flash and SenseNova models
Tokens per day
Not published
Card needed
No (mobile SMS signup)
Phone needed
Yes
Offer type
Not published
Base URL
https://token.sensenova.cn/v1
Limits
Credits per rolling 5 hours and per week (reported 120K/5h and 1.2M/week since 27 Oct); credits, not tokens
Notes
Also Anthropic Messages compatible. Credit pool figures from third-party trackers; official page is JS-only.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://token.sensenova.cn/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
30 Oct 2026
OpenAI compatibleOfficial provider
What you get free
500K tokens/month shared across built-in models (limited-time)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://ai-gateway.edgeone.link/v1
Limits
300 s max per request; rate limits on abuse
Notes
Built-in models: @makers/hy3, hy3-preview, deepseek-v4-pro, deepseek-v4-flash, deepseek-v4.1-flash, minimax-m3, minimax-m2.7, kimi-k2.6, zhuque-text. Not for production. Also BYOK for OpenAI/Anthropic/Google etc. with no platform fee. Anthropic compatible too.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://ai-gateway.edgeone.link/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
29 Oct 2026
OpenAI compatibleOfficial provider
What you get free
Free tier on Typhoon models (light usage)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.opentyphoon.ai/v1
Limits
5 req/s, 200 req/min per model
Notes
Thai-first models: typhoon-v2.5-30b-a3b-instruct, typhoon-v2.1-12b-instruct; also OCR and ASR APIs. Key from Typhoon Playground.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.opentyphoon.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
29 Oct 2026
OpenAI compatibleOfficial providerTrial / one-time
What you get free
$25 signup credit (launch offer)
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
trial
Base URL
https://api.river.ai/v1
Limits
Not published
Notes
Credit announced by CEO Igor Babuschkin on X on 11 Oct 2026 for sign-ups 'right now'; not on the pricing page, may have ended. Models include DeepSeek V4.1 Flash and GLM-5.3 Flash.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.river.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
28 Oct 2026
OpenAI compatible1M / day
What you get free
1M tokens/day shared on free core models (DeepSeek V4.1 Flash, GPT-6 Luna, Clef)
Tokens per day
1,000,000
Card needed
Yes
Phone needed
Not published
Offer type
Not published
Base URL
https://api.experientiallabs.ai/v1
Limits
10 RPM, 250K tokens/min shared
Notes
One-time $1 card check needed for free models (credited back as balance). Also Anthropic Messages and Responses API.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.experientiallabs.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
27 Oct 2026
OpenAI compatibleRelay warning
What you get free
$1 of usage per month on the Free plan
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.fluxrouter.ai/v1
Limits
Hard stop at $1/month
Notes
Smart-routing gateway (flux-auto), 78+ models, also Anthropic compatible.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.fluxrouter.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
27 Oct 2026
OpenAI compatibleRelay warning
What you get free
25 free calls, then 25 calls/day free on 10+ NVIDIA NIM open models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.inferall.ai/v1
Limits
25 calls/day (50 chat/day after one-time $5 starter pack)
Notes
Also Anthropic compatible at https://api.inferall.ai. Run by Kindly Robotics.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.inferall.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
27 Oct 2026
OpenAI compatible
What you get free
Free usage on 5 open models (DeepSeek V4.1 Flash, GLM 5.3, GLM 5.3 Flash, gpt-oss 120B, Qwen3.8 Flash)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.runcrate.ai/v1
Limits
Session cap plus weekly allowance (resets Monday); sizes unpublished
Notes
Needs a GitHub account at least 30 days old. Free requests served by third-party providers.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.runcrate.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
27 Oct 2026
OpenAI compatible7M / dayRelay warning
What you get free
7M tokens/day (Free plan, forever $0)
Tokens per day
7,000,000
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.flexlora.com/v1
Limits
Not published
Notes
Looks like a NaraRouter rebrand/reseller.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.flexlora.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
26 Oct 2026
OpenAI compatibleOfficial providerTrial / one-time
What you get free
100M one-time free tokens
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://api.inceptionlabs.ai/v1
Limits
Not published
Notes
One-time grant for new accounts.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.inceptionlabs.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
26 Oct 2026
OpenAI compatible
What you get free
30M tokens/month on Qwen3.8-27B
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.iteracompute.com/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.iteracompute.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
26 Oct 2026
OpenAI compatible10K / dayRelay warning
What you get free
10,000 tokens/day (limited time), 30+ models
Tokens per day
10,000
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.mobike.io/v1
Limits
Not published
Notes
Operator unknown.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.mobike.io/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
23 Oct 2026
OpenAI compatible10K / day
What you get free
10,000 tokens/day on free models
Tokens per day
10,000
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.euron.one/api/v1/euri
Limits
Not published
Notes
Plus plan: 100,000/day.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.euron.one/api/v1/euri",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
23 Oct 2026
OpenAI compatible1M / dayRelay warning
What you get free
$5 signup credit + 1M tokens/day on Free plan
Tokens per day
1,000,000
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.gwbus.com/v1
Limits
Not published
Notes
50+ models; new, little-known gateway.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.gwbus.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
23 Oct 2026
OpenAI compatibleRelay warning
What you get free
$0 tier with Free-labelled models
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.rout.my/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.rout.my/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
23 Oct 2026
OpenAI compatible20K / day
What you get free
20,000 tokens/day
Tokens per day
20,000
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.aionlabs.ai/v1
Limits
15 RPM
Notes
Roleplay models.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.aionlabs.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
22 Oct 2026
OpenAI: not publishedRelay warningTrial / one-time
What you get free
500K one-time trial tokens
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
trial
Base URL
Not published
Limits
Not published
Notes
Standard text models.
Source
Official page
·
Last verified
22 Oct 2026
OpenAI compatibleRelay warning
What you get free
500 requests/day
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.frenix.sh/v1
Limits
500 requests/day
Notes
Telegram sign-in.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.frenix.sh/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
22 Oct 2026
OpenAI compatibleRelay warningTrial / one-time
What you get free
$5 signup credit (one-time)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://api.mixroute.ai/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.mixroute.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
22 Oct 2026
OpenAI: not published
What you get free
$1 free credits every 4 weeks
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
Not published
Limits
Not published
Notes
Open models up to 7B.
Source
Official page
·
Last verified
21 Oct 2026
OpenAI compatibleOfficial provider
What you get free
Daily free calls
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api-inference.modelscope.cn/v1
Limits
Not published
Notes
Needs Alibaba Cloud real-name verification.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api-inference.modelscope.cn/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
21 Oct 2026
OpenAI: partial
What you get free
$26 intro credits
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.router.com/v1
Limits
Not published
Notes
Responses API only, no Chat Completions.
Source
Official page
·
Last verified
21 Oct 2026
OpenAI compatibleRelay warning
What you get free
10,000 credits ≈ $1/month
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.router9.com/v1
Limits
5 requests per 4 hours
Notes
Free-tier models only.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.router9.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
21 Oct 2026
OpenAI compatibleRelay warning
What you get free
10 requests/day on :free models (up to 200 with ads) + $0.25 weekly credit
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.electronhub.ai/v1
Limits
5 RPM
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.electronhub.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
20 Oct 2026
OpenAI compatible
What you get free
14 zero-priced limited-time free models
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://opencode.ai/inference/openai/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://opencode.ai/inference/openai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
20 Oct 2026
OpenAI compatible
What you get free
Quest Pollen (amounts unpublished)
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://gen.pollinations.ai/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://gen.pollinations.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
20 Oct 2026
OpenAI compatible
What you get free
Zero-priced free models, no credits needed
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.llmgateway.io/v1
Limits
5 requests per 10 min per model
Notes
Verified email needed.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.llmgateway.io/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
19 Oct 2026
OpenAI compatibleRelay warning
What you get free
Free models
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://onomeo.com/v1
Limits
30 calls per 5 hours per account
Notes
Beta; free answers carry a 'via onomeo.com' line.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://onomeo.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
19 Oct 2026
OpenAI compatibleRelay warning
What you get free
Free endpoint
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.blazeapi.org/free/v1/chat
Limits
300 requests/day
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.blazeapi.org/free/v1/chat",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
16 Oct 2026
OpenAI compatibleRelay warning
What you get free
79 :free models
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://crowllm.com/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://crowllm.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
16 Oct 2026
OpenAI compatible2M / day
What you get free
2M tokens/day on vikasit-nova
Tokens per day
2,000,000
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.vikasit.ai/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.vikasit.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
16 Oct 2026
OpenAI compatible500K / dayRelay warning
What you get free
500K tokens/day (1M/day with Telegram verify), 40+ free models
Tokens per day
500,000
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.xkiro.com/v1
Limits
Not published
Notes
Also a one-off 30M bonus for 30 days after Telegram verification.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.xkiro.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
16 Oct 2026
OpenAI: not published
What you get free
Not published
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
Not published
Limits
Not published
Source
Official page
·
Last verified
15 Oct 2026
OpenAI compatible1M / dayRelay warning
What you get free
63 -free models; 1M tokens/day shared after $1+ top-up
Tokens per day
1,000,000
Card needed
Top-up
Phone needed
Not published
Offer type
Not published
Base URL
https://aihubmix.com/v1
Limits
100 requests/day, 10/min
Notes
10 trial calls before top-up.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://aihubmix.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
15 Oct 2026
OpenAI compatibleOfficial provider
What you get free
Free models with a daily points cap
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://developer.amd.com.cn/radeon/api/v1
Limits
Per-minute + daily points cap (resets 00:00 Beijing)
Notes
China site only; dev/trial use, not production.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://developer.amd.com.cn/radeon/api/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
26 Oct 2026
OpenAI compatibleRelay warning
What you get free
5M tokens/month on 33 chat models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://cleanapis.com/v1
Limits
Not published
Notes
Young domain.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://cleanapis.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
15 Oct 2026
OpenAI compatibleRelay warning
What you get free
:free models
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.routeway.ai/v1
Limits
200 requests/day, 5 RPM
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.routeway.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
15 Oct 2026
OpenAI compatible
What you get free
Free credit (amount not confirmed)
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://ai-gateway.vercel.sh/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://ai-gateway.vercel.sh/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
15 Oct 2026
OpenAI compatibleOfficial provider
What you get free
8 zero-priced :free models
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://inference-api.nousresearch.com/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://inference-api.nousresearch.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
14 Oct 2026
OpenAI compatible
What you get free
5 free models + orcarouter/free router
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.orcarouter.ai/v1
Limits
10 RPM, 50 requests/day (800/day after $20 lifetime spend)
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.orcarouter.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
14 Oct 2026
OpenAI compatibleTrial / one-time
What you get free
One-time 10M tokens after email verify
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
trial
Base URL
https://api.batchrate.ai/v1
Limits
Not published
Notes
Files/Batches API only.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.batchrate.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
13 Oct 2026
OpenAI compatibleRelay warningTrial / one-time
What you get free
250K-token lifetime trial (one key)
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
trial
Base URL
https://paraloncloud.com/v1
Limits
20 RPM, 5 images/day
Notes
Unlock via X/GitHub/Discord.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://paraloncloud.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
13 Oct 2026
OpenAI compatibleTrial / one-time
What you get free
1M free tokens to start (one-time)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://api.trustedrouter.com/v1
Limits
Not published
Notes
One-time grant.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.trustedrouter.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
13 Oct 2026
OpenAI compatibleRelay warning
What you get free
:free models
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.freetheai.org/v1
Limits
50 requests/day (+50 with Discord)
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.freetheai.org/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
12 Oct 2026
OpenAI compatible
What you get free
Keyless :free models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.kilo.ai/api/gateway
Limits
200 requests/hour per IP
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.kilo.ai/api/gateway",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
12 Oct 2026
OpenAI compatibleOfficial provider
What you get free
Keyless free access
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://oai.endpoints.kepler.ai.cloud.ovh.net/v1
Limits
2 requests/min per IP per model
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://oai.endpoints.kepler.ai.cloud.ovh.net/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
12 Oct 2026
OpenAI compatible
What you get free
Free models
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://router.requesty.ai/v1
Limits
200 requests/day
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://router.requesty.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
12 Oct 2026
OpenAI compatible100K / dayRelay warning
What you get free
100K tokens/day, 500K/week
Tokens per day
100,000
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.freeaiapikey.com/v1
Limits
20 requests/day
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.freeaiapikey.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
9 Oct 2026
OpenAI compatibleRelay warning
What you get free
≥1M tokens/month on any catalogue model
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://proxy.gonkabroker.com/v1
Limits
Not published
Notes
Renews on the 1st of each month.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://proxy.gonkabroker.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
9 Oct 2026
OpenAI compatible
What you get free
:free models at $0
Tokens per day
Not published
Card needed
Top-up
Phone needed
Not published
Offer type
Not published
Base URL
https://llm.onerouter.pro/v1
Limits
1,000 requests/day shared across free models
Notes
Standard free models need a $5+ balance (not consumed).
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://llm.onerouter.pro/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
26 Oct 2026
OpenAI compatible
What you get free
Rolling 7-day value-based allowance on :free models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://tokenharbor.ai/v1
Limits
60 requests/min, 1,800/hour (free)
Notes
Free-model prompts may be retained. Also Anthropic compatible.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://tokenharbor.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
26 Oct 2026
OpenAI compatibleRelay warning
What you get free
No key needed; hourly fair use
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://vireonix.ai/v1
Limits
Hourly fair-use cap
Notes
Keyless.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://vireonix.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
9 Oct 2026
OpenAI compatible100K / dayRelay warning
What you get free
100K ANY Tokens/day on free models
Tokens per day
100,000
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.anyapi.ai/v1
Limits
Not published
Notes
Allowance is in ANY Tokens.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.anyapi.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
8 Oct 2026
OpenAI compatible500K / dayRelay warning
What you get free
500,000 tokens/day
Tokens per day
500,000
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://dreamprompting.com/api/v1
Limits
5,000 requests/day
Notes
Still 500K/day on 16 Oct.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://dreamprompting.com/api/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
16 Oct 2026
OpenAI compatible100K / day
What you get free
100,000 tokens/24h with free token
Tokens per day
100,000
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.llm7.io/v1
Limits
Not published
Notes
Still 100K/24h on 16 Oct.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.llm7.io/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
16 Oct 2026
OpenAI compatibleRelay warning
What you get free
$20 shared pool/day plus $1 personal
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.y-api.bestvirtualgoods.com/v1
Limits
Not published
Notes
Shared public key; reseller.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.y-api.bestvirtualgoods.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
8 Oct 2026
OpenAI compatibleRelay warningTrial / one-time
What you get free
Signup trial credits + 6 free models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://api.bazaarlink.ai/v1
Limits
10 RPM / 60 units
Notes
Free list grew from 2 to 7 models on 20 Oct, now 6 (28 Oct). Limits shared across free models, resets 00:00 UTC.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.bazaarlink.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
28 Oct 2026
OpenAI compatibleRelay warning
What you get free
Daily token lucky draw (amount not published)
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
Not published
Limits
Not published
Notes
Third-party relay; genuinely free models mainly MiMo and MiniMax. Also Anthropic compatible.
Source
Official page
·
Last verified
7 Oct 2026
OpenAI compatibleRelay warning
What you get free
Daily check-in credit (size not stated)
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.wanapis.com/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.wanapis.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
7 Oct 2026
OpenAI compatibleRelay warningTrial / one-time
What you get free
$5 one-time signup credit
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://api.yepapi.com/v1/ai
Limits
Not published
Notes
One-time credit.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.yepapi.com/v1/ai",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
7 Oct 2026
OpenAI compatible30K / day
What you get free
30,000 tokens/day on self-hosted models
Tokens per day
30,000
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.free.ai/v1
Limits
1,000 calls/month, 60/min
Notes
Self-hosted models like Qwen 2.5.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.free.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
6 Oct 2026
OpenAI compatibleRelay warning
What you get free
105 :free models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.unorouter.com/v1
Limits
1 request/min per free model
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.unorouter.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
6 Oct 2026
OpenAI compatible7M / dayRelay warning
What you get free
7M tokens/day on 1 free model
Tokens per day
7,000,000
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://router.bynara.id/v1
Limits
15 requests/min
Notes
Free plan now only laguna-s-2.1 (cut 10→5 on 14 Oct, 5→3 on 26 Oct, 3→1 on 28 Oct). Also Anthropic compatible.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://router.bynara.id/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
28 Oct 2026
OpenAI compatibleOfficial providerTrial / one-time
What you get free
Free trial on hosted models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://integrate.api.nvidia.com/v1
Limits
40 requests/min, 10,000 requests/day
Notes
Models include Kimi K3, DeepSeek V4.1 Flash, GLM 5.3.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://integrate.api.nvidia.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
5 Oct 2026
OpenAI compatibleOfficial provider
What you get free
Free quota for new users
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://dashscope-intl.aliyuncs.com/compatible-mode/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://dashscope-intl.aliyuncs.com/compatible-mode/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Oct 2026
OpenAI compatibleOfficial provider
What you get free
Not published
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.anthropic.com/v1/
Limits
Not published
Notes
OpenAI SDK compatibility layer; no free daily allowance published.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.anthropic.com/v1/",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Oct 2026
OpenAI compatible3M / dayRelay warning
What you get free
3M tokens/day on free-category models
Tokens per day
3,000,000
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.apinex.bond/v1
Limits
5 requests/min per IP (free)
Notes
Cut from 10M to 3M/day on 23 Oct. Free models: free/gpt-6-luna, free/deepseek-v4.1-flash, free/gemini-3.8-flash. Paid 10/20/30-day plans. Model names are its own catalogue labels.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.apinex.bond/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
23 Oct 2026
OpenAI compatible1M / dayOfficial providerTrial / one-time
What you get free
1,000,000 tokens/day during $5 30-day trial
Tokens per day
1,000,000
Card needed
Yes
Phone needed
Not published
Offer type
trial
Base URL
https://api.cerebras.ai/v1
Limits
Not published
Notes
Trial only, card required, 30 days.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.cerebras.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Oct 2026
OpenAI compatibleOfficial provider
What you get free
10,000 Neurons/day (not tokens)
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.cloudflare.com/client/v4/accounts/{ACCOUNT_ID}/ai/v1
Limits
Not published
Notes
Allowance is in Neurons; token equivalent depends on model.
Source
Official page
·
Last verified
3 Oct 2026
OpenAI compatibleOfficial providerTrial / one-time
What you get free
Trial key: 1,000 calls/month
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
trial
Base URL
https://api.cohere.ai/compatibility/v1
Limits
1,000 calls/month
Notes
Trial keys are for non-production use.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.cohere.ai/compatibility/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Oct 2026
OpenAI compatibleOfficial provider
What you get free
Free tier on selected Gemini models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://generativelanguage.googleapis.com/v1beta/openai/
Limits
Per-model RPM/RPD on Google's rate-limit page
Notes
Free-tier prompts may be used to improve Google products.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://generativelanguage.googleapis.com/v1beta/openai/",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Oct 2026
Not OpenAI compatibleOfficial provider
What you get free
90-day free use of Gemini models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
Not published
Limits
Not published
Notes
Not OpenAI compatible (Vertex/Gemini API).
Source
Official page
·
Last verified
3 Oct 2026
OpenAI compatible200K / dayOfficial provider
What you get free
200,000 tokens/day on listed models
Tokens per day
200,000
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://api.groq.com/openai/v1
Limits
1,000 requests/day on listed models
Notes
Limits vary by model.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.groq.com/openai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Oct 2026
OpenAI compatibleOfficial provider
What you get free
Free experiment tier
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.mistral.ai/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.mistral.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Oct 2026
OpenAI compatible
What you get free
:free models
Tokens per day
Not published
Card needed
No
Phone needed
Not published
Offer type
Not published
Base URL
https://openrouter.ai/api/v1
Limits
50 requests/day until credits are bought
Notes
Free-model providers may log prompts.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://openrouter.ai/api/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Oct 2026
OpenAI compatible200K / dayOfficial provider
What you get free
200,000 tokens/day per model (as of 3 Oct)
Tokens per day
200,000
Card needed
Yes
Phone needed
Not published
Offer type
Not published
Base URL
https://api.sambanova.ai/v1
Limits
20 requests/day
Notes
Since 15 Oct the Free plan requires a card and a purchase.
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.sambanova.ai/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
15 Oct 2026
OpenAI compatibleOfficial provider
What you get free
Free models
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.siliconflow.com/v1
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.siliconflow.com/v1",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Oct 2026
OpenAI compatibleOfficial provider
What you get free
Free Flash models
Tokens per day
Not published
Card needed
Not published
Phone needed
Not published
Offer type
Not published
Base URL
https://api.z.ai/api/paas/v4
Limits
Not published
Copy Python starter
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["API_KEY"],
    base_url="https://api.z.ai/api/paas/v4",
)

resp = client.chat.completions.create(
    model="MODEL_ID",  # copy the model id from the provider's docs
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Copy the model id from the provider’s official docs. Do not guess it.

Source
Official page
·
Last verified
3 Oct 2026
Full snapshot table (all 124 providers)

This table is the complete 14 Nov 2026 snapshot, including ended offers. Figures are copied from each provider’s own pages. “Not published” means no official number was found.

Free and free-tier LLM API providers, snapshot 14 Nov 2026, 03:44 UTC.
ProviderStatusOpenAI compatibleFree allowanceTokens/dayCardPhoneOffer typeBase URLRelay warningLast verified
1vyActiveYes$1 free credit on signup, 100+ models on one key (15% markup after)Not publishedNot publishedNot publishedNot publishedhttps://api.1vy.ai/v1Yes11 Nov 2026
AevonActiveYes$100 free credits on signup via GitHub; 3 Free Tier modelsNot publishedNoNot publishedNot publishedhttps://aevon.sh/v1Yes30 Oct 2026
Agnes AIActiveNot publishedNot publishedNot publishedNot publishedNot publishedNot publishedNot publishedNo15 Oct 2026
AIHubMixActiveYes63 -free models; 1M tokens/day shared after $1+ top-up1,000,000Top-upNot publishedNot publishedhttps://aihubmix.com/v1Yes15 Oct 2026
Aion LabsActiveYes20,000 tokens/day20,000NoNot publishedNot publishedhttps://api.aionlabs.ai/v1No22 Oct 2026
Alibaba Cloud Model StudioActiveYesFree quota for new usersNot publishedNot publishedNot publishedNot publishedhttps://dashscope-intl.aliyuncs.com/compatible-mode/v1No3 Oct 2026
AMD Radeon Cloud Token FactoryActiveYesFree models with a daily points capNot publishedNot publishedNot publishedNot publishedhttps://developer.amd.com.cn/radeon/api/v1No26 Oct 2026
AMD Token LabActiveYesComplimentary Token Lab points for free AMD AI Developer Program members (10 points, 30-day access per AMD dev community)Not publishedNoNot publishedNot publishedhttps://api.tokenlab.amd.com/v1No14 Nov 2026
Anthropic APIActiveYesNot publishedNot publishedNot publishedNot publishedNot publishedhttps://api.anthropic.com/v1/No3 Oct 2026
AnyAPIActiveYes100K ANY Tokens/day on free models100,000Not publishedNot publishedNot publishedhttps://api.anyapi.ai/v1Yes8 Oct 2026
Api.AirforceActiveYesFree plan $0/month on basic models (e.g. gpt-oss-20b)Not publishedNoNot publishedNot publishedhttps://api.airforce/v1Yes30 Oct 2026
APIBridgeActiveYes10,000 tokens/day (limited time), 30+ models10,000Not publishedNot publishedNot publishedhttps://api.mobike.io/v1Yes23 Oct 2026
APInexActiveYes3M tokens/day on free-category models3,000,000NoNot publishedNot publishedhttps://api.apinex.bond/v1Yes23 Oct 2026
ArcaneAPIActiveYes$10 free credit on signup, 53 modelsNot publishedNoNot publishedNot publishedNot publishedYes11 Nov 2026
BatchrateActiveYesOne-time 10M tokens after email verifyNot publishedNot publishedNot publishedtrialhttps://api.batchrate.ai/v1No13 Oct 2026
Bay of Assets / RouterivaActiveNot published500K one-time trial tokensNot publishedNot publishedNot publishedtrialNot publishedYes22 Oct 2026
BlazeAPIActiveYesFree endpointNot publishedNot publishedNot publishedNot publishedhttps://api.blazeapi.org/free/v1/chatYes16 Oct 2026
Boundless APIActiveYes$5 signup bonus (Starter model pool); may expire after 30 days if unusedNot publishedNoNot publishedNot publishedhttps://api.boundlessapi.com/v1Yes6 Nov 2026
BraintrustActiveYesStarter plan is $0/month and includes $10 of model credits each month for built-in modelsNot publishedNoNot publishedongoinghttps://api.braintrust.dev/v1/proxyNo13 Nov 2026
BytezActiveNot published$1 free credits every 4 weeksNot publishedNot publishedNot publishedNot publishedNot publishedNo21 Oct 2026
CerebrasActiveYes1,000,000 tokens/day during $5 30-day trial1,000,000YesNot publishedtrialhttps://api.cerebras.ai/v1No3 Oct 2026
Clean APIsActiveYes5M tokens/month on 33 chat modelsNot publishedNoNot publishedNot publishedhttps://cleanapis.com/v1Yes15 Oct 2026
Cloudflare Workers AIActiveYes10,000 Neurons/day (not tokens)Not publishedNoNot publishedNot publishedhttps://api.cloudflare.com/client/v4/accounts/{ACCOUNT_ID}/ai/v1No3 Oct 2026
CohereActiveYesTrial key: 1,000 calls/monthNot publishedNoNot publishedtrialhttps://api.cohere.ai/compatibility/v1No3 Oct 2026
CrowLLMActiveYes79 :free modelsNot publishedNot publishedNot publishedNot publishedhttps://crowllm.com/v1Yes16 Oct 2026
DreamPromptingActiveYes500,000 tokens/day500,000NoNot publishedNot publishedhttps://dreamprompting.com/api/v1Yes16 Oct 2026
Electron HubActiveYes10 requests/day on :free models (up to 200 with ads) + $0.25 weekly creditNot publishedNot publishedNot publishedNot publishedhttps://api.electronhub.ai/v1Yes20 Oct 2026
EURI by EuronActiveYes10,000 tokens/day on free models10,000Not publishedNot publishedNot publishedhttps://api.euron.one/api/v1/euriNo23 Oct 2026
EXAONE API Platform (LG K-EXAONE 2.5)ActiveYesK-EXAONE 2.5 API free to use with an API key (announced 9 Nov 2026); no quota publishedNot publishedNot publishedNot publishedNot publishedhttps://api.exaone.ai/v1No9 Nov 2026
Experiential LabsActiveYes1M tokens/day shared on free core models (DeepSeek V4.1 Flash, GPT-6 Luna, Clef)1,000,000YesNot publishedNot publishedhttps://api.experientiallabs.ai/v1No27 Oct 2026
FlexloraActiveYes7M tokens/day (Free plan, forever $0)7,000,000NoNot publishedNot publishedhttps://api.flexlora.com/v1Yes26 Oct 2026
Flux RouterActiveYes$1 of usage per month on the Free planNot publishedNot publishedNot publishedNot publishedhttps://api.fluxrouter.ai/v1Yes27 Oct 2026
Free.aiActiveYes30,000 tokens/day on self-hosted models30,000NoNot publishedNot publishedhttps://api.free.ai/v1No6 Oct 2026
FreeAIapikeyActiveYes100K tokens/day, 500K/week100,000Not publishedNot publishedNot publishedhttps://api.freeaiapikey.com/v1Yes9 Oct 2026
FreeInference (Harvard MadSys Lab)ActiveYesFree for research/education, 'generous' quota, tokens/day unpublishedNot publishedNoNot publishedNot publishedhttps://freeinference.org/v1No10 Nov 2026
FreeTheAIActiveYes:free modelsNot publishedNot publishedNot publishedNot publishedhttps://api.freetheai.org/v1Yes12 Oct 2026
FrenixActiveYes500 requests/dayNot publishedNot publishedNot publishedNot publishedhttps://api.frenix.sh/v1Yes22 Oct 2026
Gemini API (Google AI Studio)ActiveYesFree tier on selected Gemini modelsNot publishedNoNot publishedNot publishedhttps://generativelanguage.googleapis.com/v1beta/openai/No3 Oct 2026
Gonka BrokerActiveYes≥1M tokens/month on any catalogue modelNot publishedNot publishedNot publishedNot publishedhttps://proxy.gonkabroker.com/v1Yes9 Oct 2026
GonkaRouterActiveYes1M free tokens one-time for new accounts (email or Google sign-in)Not publishedNoNot publishedtrialhttps://api.gonkarouter.io/v1Yes13 Nov 2026
Google Cloud Vertex AI express modeActiveNo90-day free use of Gemini modelsNot publishedNoNot publishedNot publishedNot publishedNo3 Oct 2026
GroqActiveYes200,000 tokens/day on listed models200,000NoNot publishedNot publishedhttps://api.groq.com/openai/v1No3 Oct 2026
GW BUSActiveYes$5 signup credit + 1M tokens/day on Free plan1,000,000Not publishedNot publishedNot publishedhttps://api.gwbus.com/v1Yes23 Oct 2026
Helyx AIActiveYes$10 signup credit (one-time) on any listed modelNot publishedNoNot publishedtrialhttps://helyxai.space/v1Yes2 Nov 2026
Hetzner Inference API (Experiments)ActiveYesFree of charge while the API is experimental; models Qwen3.6-35B-A3B-FP8 and Qwen3.8-27BNot publishedNot publishedNot publishedNot publishedhttps://inference.hetzner.com/api/v1No5 Nov 2026
HyperbolicActiveYesOne-time $1 credit after phone verification (unlocks inference on Free Tier); credits never expire once boughtNot publishedNoYestrialhttps://api.hyperbolic.ai/inference/v1No11 Nov 2026
IBM watsonx.aiActiveNot publishedFree plan: up to 300,000 foundation-model tokens per month, up to 20 CUH of machine learning tools, and up to 100 documents for text extractionNot publishedNot publishedNot publishedongoinghttps://us-south.ml.cloud.ibm.comNo13 Nov 2026
Inception (Mercury)ActiveYes100M one-time free tokensNot publishedNoNot publishedtrialhttps://api.inceptionlabs.ai/v1No26 Oct 2026
InferAllActiveYes25 free calls, then 25 calls/day free on 10+ NVIDIA NIM open modelsNot publishedNoNot publishedNot publishedhttps://api.inferall.ai/v1Yes27 Oct 2026
InferbaseActiveYesFree plan: 5,000 requests/month, 100 managed requests and $0.50 model usage/day on models up to $2/M output tokensNot publishedNoNot publishedNot publishedNot publishedYes10 Nov 2026
Infercom (SambaNova EU)ActiveYes€5 free API credit on signup (expires after 90 days)Not publishedNoNot publishedNot publishedhttps://api.infercom.ai/v1No2 Nov 2026
InfronActiveYes:free models at $0Not publishedTop-upNot publishedNot publishedhttps://llm.onerouter.pro/v1No26 Oct 2026
IteraComputeActiveYes30M tokens/month on Qwen3.8-27BNot publishedNoNot publishedNot publishedhttps://api.iteracompute.com/v1No26 Oct 2026
JoinGonka GatewayActiveYes3M one-time starter tokens after email confirmationNot publishedNo cardNot publishedtrialhttps://gate.joingonka.ai/v1Yes3 Nov 2026
Kilo GatewayActiveYesKeyless :free modelsNot publishedNoNot publishedNot publishedhttps://api.kilo.ai/api/gatewayNo12 Oct 2026
Lightning AI Model APIsActiveYes5 free credits on signup (25 more if you add a card); docs say up to 40M free tokens to start; credits expire after 12 monthsNot publishedNoNot publishedNot publishedhttps://lightning.ai/api/v1No5 Nov 2026
Liquid AIActiveYesLimited-time $2 in free API credits for new accounts (stated on console.liquid.ai)Not publishedNot publishedNot publishedtrialhttps://api.liquid.ai/openai/v1No13 Nov 2026
LLM EUActiveYes$5 promotional inference credit per new organisation; expires in 90 days; one grant per orgNot publishedNoNot publishedNot publishedhttps://api.llmeu.com/v1Yes6 Nov 2026
LLM GatewayActiveYesZero-priced free models, no credits neededNot publishedNot publishedNot publishedNot publishedhttps://api.llmgateway.io/v1No19 Oct 2026
LLM7.ioActiveYes100,000 tokens/24h with free token100,000NoNot publishedNot publishedhttps://api.llm7.io/v1No16 Oct 2026
LLMCloudActiveYes$1 free credits on signup; no card requiredNot publishedNoNot publishedtrialhttps://api.llmcloud.ai/v1No6 Nov 2026
Mango InferenceActiveYes$10 free credits on signup (launch offer); 30% off models through 11 Nov 2026Not publishedNoNot publishedNot publishedhttps://api.mangoboost.io/v1No6 Nov 2026
Maritaca AI (S5-free)ActiveYesS5-free model: no token charge for a limited launch window (text only; no Batch/Flex/built-in tools)Not publishedNoNot publishedNot publishedhttps://chat.maritaca.ai/apiNo6 Nov 2026
MervexActiveYesFree key: 100 requests/day plus one-time +400 welcome bonus (per the team's 3 Nov launch post)Not publishedNo cardNot publishedNot publishedhttps://api.mervex.cc.cd/v1Yes4 Nov 2026
Mistral (La Plateforme)ActiveYesFree experiment tierNot publishedNot publishedNot publishedNot publishedhttps://api.mistral.ai/v1No3 Oct 2026
MixRouteActiveYes$5 signup credit (one-time)Not publishedNoNot publishedtrialhttps://api.mixroute.ai/v1Yes22 Oct 2026
ModelScope API-InferenceActiveYesDaily free callsNot publishedNot publishedNot publishedNot publishedhttps://api-inference.modelscope.cn/v1No21 Oct 2026
nabh.cloudActiveYesFree plan: 50 credits (₹50) on signup, then 50 every month; GPT OSS 20B onlyNot publishedNo card neededNot publishedNot publishedhttps://api.nabh.cloud/v1No4 Nov 2026
NaraRouterActiveYes7M tokens/day on 1 free model7,000,000NoNot publishedNot publishedhttps://router.bynara.id/v1Yes28 Oct 2026
NFlashXActiveYes$1 welcome credit on signup (first 100 accounts) plus free-tier Flash modelNot publishedNoNot publishedNot publishedhttps://nflashx.cc.cd/v1Yes14 Nov 2026
ngrok.ai AI GatewayActiveYes$1 credit for new accountsNot publishedNo card stated for the $1 creditNot publishedtrialhttps://gateway.ngrok.ai/v1No3 Nov 2026
NLP CloudActiveNoFree plan can test the I models without a credit card. The pay-as-you-go plan adds an initial $15 credit and requires a cardNot publishedNoNot publishedongoinghttps://api.nlpcloud.io/v1No13 Nov 2026
Nous PortalActiveYes8 zero-priced :free modelsNot publishedNot publishedNot publishedNot publishedhttps://inference-api.nousresearch.com/v1No14 Oct 2026
NovAIActiveYes$2 signup credit after email verify; 5 free Zhipu Flash models (glm-4.7-flash etc.)Not publishedNoNot publishedtrialhttps://aiapi-pro.com/v1Yes12 Nov 2026
Novita AIActiveYesApodex 1.1 Mini is listed with Free input and Free output on the official pricing pageNot publishedNot publishedNot publishedongoinghttps://api.novita.ai/openai/v1No13 Nov 2026
NscaleActiveYes$5 free starter credit for new accounts (stated in docs changelog for Serverless Inference and Fine-tuning); quickstart says early users may be eligibleNot publishedNoNot publishedNot publishedhttps://inference.api.nscale.com/v1No6 Nov 2026
NVIDIA Build / NIMActiveYesFree trial on hosted modelsNot publishedNoNot publishedtrialhttps://integrate.api.nvidia.com/v1No5 Oct 2026
OnomeoActiveYesFree modelsNot publishedNot publishedNot publishedNot publishedhttps://onomeo.com/v1Yes19 Oct 2026
OpenAnyActiveYes$0.50 signup credit (homepage; launch post said $1), does not expireNot publishedNo card for the signup creditNot publishedtrialhttps://api.openany.si/v1Yes4 Nov 2026
OpenCode ZenActiveYes14 zero-priced limited-time free modelsNot publishedNot publishedNot publishedNot publishedhttps://opencode.ai/inference/openai/v1No20 Oct 2026
OpenPathsActiveYesFree models bill $0 per token (OpenRouter :free routes plus OpenPaths Free and others); paid models need creditsNot publishedNoNot publishedNot publishedhttps://openpaths.io/v1Yes6 Nov 2026
OpenRouterActiveYes:free modelsNot publishedNoNot publishedNot publishedhttps://openrouter.ai/api/v1No3 Oct 2026
OrcaRouterActiveYes5 free models + orcarouter/free routerNot publishedNoNot publishedNot publishedhttps://api.orcarouter.ai/v1No14 Oct 2026
Orq AI GatewayActiveYes$1 free credit on signupNot publishedNo cardNot publishedNot publishedhttps://my.orq.ai/v3/routerNo4 Nov 2026
OVHcloud AI EndpointsActiveYesKeyless free accessNot publishedNoNot publishedNot publishedhttps://oai.endpoints.kepler.ai.cloud.ovh.net/v1No12 Oct 2026
ParalonCloudActiveYes250K-token lifetime trial (one key)Not publishedNot publishedNot publishedtrialhttps://paraloncloud.com/v1Yes13 Oct 2026
PEKPIK LLMActiveYes$1 on email verify, then one daily claim (up to $20 over first 40 days, continues after)Not publishedNoNot publishedNot publishedhttps://aiapiv2.pekpik.com/v1Yes30 Oct 2026
PollinationsActiveYesQuest Pollen (amounts unpublished)Not publishedNot publishedNot publishedNot publishedhttps://gen.pollinations.ai/v1No20 Oct 2026
Qora APIActiveYes$5 trial credit on request (email support or DM @QoraAPI); not automaticNot publishedNot publishedNot publishedNot publishedhttps://qoraapi.com/v1Yes14 Nov 2026
Ramp RouterActivePartial$26 intro creditsNot publishedNot publishedNot publishedNot publishedhttps://api.router.com/v1No21 Oct 2026
Regolo.aiActiveYes30-day free trial, 1M tokens/day on all Core models1,000,000NoNot publishedtrialhttps://api.regolo.ai/v1No9 Nov 2026
RequestyActiveYesFree modelsNot publishedNot publishedNot publishedNot publishedhttps://router.requesty.ai/v1No12 Oct 2026
River AI (River Inference)ActiveYes$25 signup credit (launch offer)Not publishedNot publishedNot publishedtrialhttps://api.river.ai/v1No28 Oct 2026
rout.myActiveYes$0 tier with Free-labelled modelsNot publishedNot publishedNot publishedNot publishedhttps://api.rout.my/v1Yes23 Oct 2026
Router9ActiveYes10,000 credits ≈ $1/monthNot publishedNot publishedNot publishedNot publishedhttps://api.router9.com/v1Yes21 Oct 2026
RoutewayActiveYes:free modelsNot publishedNot publishedNot publishedNot publishedhttps://api.routeway.ai/v1Yes15 Oct 2026
RuncrateActiveYesFree usage on 5 open models (DeepSeek V4.1 Flash, GLM 5.3, GLM 5.3 Flash, gpt-oss 120B, Qwen3.8 Flash)Not publishedNoNot publishedNot publishedhttps://api.runcrate.ai/v1No27 Oct 2026
SambaNova CloudActiveYes200,000 tokens/day per model (as of 3 Oct)200,000YesNot publishedNot publishedhttps://api.sambanova.ai/v1No15 Oct 2026
SenseNova Token PlanActiveYesFree public-beta Token Plan: shared credit pool across Kimi K3, GLM-5.2, DeepSeek V4.1 Flash and SenseNova modelsNot publishedNo (mobile SMS signup)YesNot publishedhttps://token.sensenova.cn/v1No30 Oct 2026
SiliconFlowActiveYesFree modelsNot publishedNot publishedNot publishedNot publishedhttps://api.siliconflow.com/v1No3 Oct 2026
SpekaActiveYesFree plan includes $1 of model usage each month and access to all open catalog modelsNot publishedNoNot publishedongoinghttps://speka.me/v1Yes13 Nov 2026
Tencent Cloud TokenHubActiveNot published1M free tokens per language model, one-time, valid 1 year (campaign until 31 Dec 2026)Not publishedTop-upNot publishedtrialNot publishedNo2 Nov 2026
Tencent EdgeOne Makers ModelsActiveYes500K tokens/month shared across built-in models (limited-time)Not publishedNoNot publishedNot publishedhttps://ai-gateway.edgeone.link/v1No29 Oct 2026
Together AIActiveYesTernary Bonsai 27B (Prism-ML/Ternary-Bonsai-27B) is listed as Free for input and output on the serverless models pageNot publishedNot publishedNot publishedongoinghttps://api.together.ai/v1No13 Nov 2026
Token HarborActiveYesRolling 7-day value-based allowance on :free modelsNot publishedNoNot publishedNot publishedhttps://tokenharbor.ai/v1No26 Oct 2026
TokenFaucetActiveYesDaily token lucky draw (amount not published)Not publishedNot publishedNot publishedNot publishedNot publishedYes7 Oct 2026
TrustedRouterActiveYes1M free tokens to start (one-time)Not publishedNoNot publishedtrialhttps://api.trustedrouter.com/v1No13 Oct 2026
Typhoon API (OpenTyphoon)ActiveYesFree tier on Typhoon models (light usage)Not publishedNoNot publishedNot publishedhttps://api.opentyphoon.ai/v1No29 Oct 2026
UnoRouterActiveYes105 :free modelsNot publishedNoNot publishedNot publishedhttps://api.unorouter.com/v1Yes6 Oct 2026
VectoreeActiveYes$5 one-time welcome credit after CLI login and project link; then pay-as-you-go walletNot publishedNoNot publishedtrialhttps://vectoree.ai/api/v1Yes13 Nov 2026
Vercel AI GatewayActiveYesFree credit (amount not confirmed)Not publishedNot publishedNot publishedNot publishedhttps://ai-gateway.vercel.sh/v1No15 Oct 2026
Vikasit AIActiveYes2M tokens/day on vikasit-nova2,000,000Not publishedNot publishedNot publishedhttps://api.vikasit.ai/v1No16 Oct 2026
VireonixActiveYesNo key needed; hourly fair useNot publishedNoNot publishedNot publishedhttps://vireonix.ai/v1Yes9 Oct 2026
WanAPIsActiveYesDaily check-in credit (size not stated)Not publishedNot publishedNot publishedNot publishedhttps://api.wanapis.com/v1Yes7 Oct 2026
xKiroActiveYes500K tokens/day (1M/day with Telegram verify), 40+ free models500,000Not publishedNot publishedNot publishedhttps://api.xkiro.com/v1Yes16 Oct 2026
Y-API daily shared keyActiveYes$20 shared pool/day plus $1 personalNot publishedNot publishedNot publishedNot publishedhttps://api.y-api.bestvirtualgoods.com/v1Yes8 Oct 2026
YepAPIActiveYes$5 one-time signup creditNot publishedNoNot publishedtrialhttps://api.yepapi.com/v1/aiYes7 Oct 2026
Z.aiActiveYesFree Flash modelsNot publishedNot publishedNot publishedNot publishedhttps://api.z.ai/api/paas/v4No3 Oct 2026
Zro (MoonMath)ActiveYesFree plan: daily usage on DeepSeek V4.1 Flash only (GLM-5.3 Flash no longer free)Not publishedNot publishedNot publishedNot publishedhttps://zro.moonmath.ai/v1No12 Nov 2026
AI21 Labs (Jamba)EndedPartialfirst-party Jamba API and AI21 Studio sunset on 9 Aug 2026; the trial no longer appliesNot publishedNoNot publishedNot publishedhttps://api.ai21.com/studio/v1No11 Nov 2026
Fireworks AIEndedYesNone (signup credit removed)Not publishedNot publishedNot publishedNot publishedhttps://api.fireworks.ai/inference/v1No15 Oct 2026
Hugging Face Inference ProvidersEndedYesNone (free monthly credits ended)Not publishedNoNot publishedNot publishedhttps://router.huggingface.co/v1No15 Oct 2026
Ollama CloudEndedYesNo free cloud tier: pay-as-you-go per token ($0 plan); local use freeNot publishedNot publishedNot publishedNot publishedhttps://ollama.com/v1No10 Nov 2026

Where can I use these models for free?

A page for each model family that at least two providers in our data name. We only list a provider when that family already appears in its entry — we do not guess the rest of a catalogue.

All model pages →

Guides

Short explainers on OpenAI-compatible APIs, safety, rate limits and individual providers. Every figure is taken from the list above or from official docs — never invented.

Browse all guides →

Common questions

What does OpenAI compatible mean?

You can use the official OpenAI SDK or any client that talks to OpenAI’s API. Just change the base URL and use the provider’s API key. “Partial” means only some features or models work that way.

What are tokens per day?

A token is a small chunk of text the model reads or writes — a few characters or a short word. The daily figure is the provider’s published free limit, when they publish one. If they do not, we show “Not published” rather than guessing.

What does the relay warning mean?

Those sites are unofficial gateways or resellers. They forward your prompts to other providers. Do not send private or sensitive data through them.

Do I need a credit card?

It depends on the provider. Use the “No card needed” filter to narrow the list, then confirm on their site — some ask for a card later, or after a trial credit runs out.

Do I need a phone number?

Only some providers say so on their official pages. Use the “No phone needed” filter to see entries we could confirm do not require a phone. If the field says “Not published”, check their sign-up page — we do not guess.

What is the difference between ongoing free and a trial?

“Ongoing free” is a renewing free tier. “Trial / one-time” is a signup credit or a time-limited allowance. We only set this when official pages make the type clear; otherwise it stays “Not published”.

Why does it say “Not published”?

The provider has not published that figure on their public pages. We never invent quotas, prices or endpoints.

How often is this list updated?

Every day. The “last checked” time at the top is when the list was last reviewed, shown in UTC.