Free LLM API List (2026): Which Providers Are Still Free and How to Get a Key
Short answer: as of 31 Oct 2026, the Google Gemini API, Groq, OpenRouter's :free models, Mistral's Free mode and SentioGate's free community pool offer free LLM API access. Cerebras offers a trial with credits instead of a free tier, Hugging Face Inference Providers gives free accounts no monthly credits, and GitHub Models has been retired.
If you want a free LLM API key today, most of these let you create one without a credit card; the no-credit-card section lists which. Every row below was checked on the provider's own page on 31 Oct 2026 and links to it. Where the provider does not publish a number on that page, we say so instead of guessing. For how to build a chatbot on one of these, see the free chatbot API guide; for a working free key in a minute, see the free chatbot API quickstart.
As of 2026-10-31 (checked Oct 2026). Free tiers change often; confirm on the linked page before you rely on a limit.
Free LLM API providers in 2026
| Provider | What is free | Limits (as published) | Card needed to start? | Source (checked Oct 2026) |
|---|---|---|---|---|
| Google Gemini API | Free tier with limited access to certain models; free input and output tokens; Google AI Studio use is free | RPM, TPM and RPD per project, not per key; RPD resets at midnight Pacific; your active limits are shown in AI Studio | No: the Free tier needs an active project; billing is needed for Tier 1 | Pricing, rate limits |
| Groq | Free plan at $0 | Organization-level limits; exact numbers are on your account's limits page | No: a payment method is needed to upgrade to the Developer plan | Rate limits, billing FAQs |
OpenRouter :free models | Model variants ending in :free | 20 requests per minute; 50 per day with under 10 credits bought, 1,000 per day with 10 or more | No for the 50-per-day tier | Limits |
| Mistral | Free mode: create API keys and use included monthly usage | Per model, in tokens per minute and requests per second; numbers are on your Limits page | Not stated on the page we checked | Usage limits |
| SentioGate free community pool | Send model: "auto" with no provider keys | 600 requests per 5 hours per user; 429 community_pool_exhausted with Retry-After when used up | No | Free chatbot API, pricing |
We build SentioGate. Its free community pool is a shared community evaluation tier for prototypes, not production traffic.
Free LLM APIs with no credit card
Going by the "Card needed to start?" column above:
- No card to start: the Gemini API Free tier (billing is only needed to move to Tier 1), Groq's Free plan (a payment method is only needed to upgrade), OpenRouter's
:freemodels at the 50-requests-per-day level, and SentioGate's free community pool. - Not stated: Mistral's usage-limits page does not say whether Free mode needs a card, so check at sign-up.
- Card required: the Cerebras Free Trial gives its credits only after you add a verified payment method (see the trial section below).
How to get a free LLM API key
The steps are much the same everywhere: sign up, create a key in the provider's console, keep it in an environment variable, and point an OpenAI-compatible client at the provider's base URL.
- Google Gemini API: create a key in Google AI Studio. Limits apply per Google Cloud project, so a second key in the same project adds no capacity.
- Groq: create a key in the GroqCloud console. Your exact limits are on your account's limits page.
- OpenRouter: any OpenRouter API key can call model IDs that end in
:free. The OpenRouter free models guide shows how to list them and check your key's free usage. - Mistral: Free mode lets you create API keys and use the included monthly usage.
- SentioGate: sign up, create a gateway key in the Vault and send
model: "auto". With no provider keys, requests go to the free community pool; see the free chatbot API quickstart.
Never put a key in client-side code or a public repository. Anyone who has it can spend your free quota, or your money once billing is on.
Trial credits, not a free tier
| Provider | What you get | Source (checked Oct 2026) |
|---|---|---|
| Cerebras | No permanently free tier. A Free Trial gives $5 of credits after you add a verified payment method; the credits expire after 30 days. Trial limits for gpt-oss-120b are 5 requests per minute and 1M tokens per day. | Rate limits |
| Hugging Face Inference Providers | Free accounts get no monthly credits and must buy credits; PRO accounts get $2 of credits a month. Provider rates are passed through with no markup. | Pricing |
No longer available
| Service | Status | Source (checked Oct 2026) |
|---|---|---|
| GitHub Models | GitHub's documentation says GitHub Models was retired as of July 30, 2026. | GitHub Docs |
Things to check before you pick one
- Data use. Google says content sent on the Gemini Free tier is used to improve its products. Read each provider's terms before sending user data.
- Limit scope. Gemini limits apply per project and Groq's per organization, so extra keys do not add capacity. OpenRouter says extra accounts or keys do not raise its limits either.
- Balance rules. OpenRouter notes that a negative credit balance can cause 402 errors even on free models.
- Errors. Groq and SentioGate return 429 with a retry-after header when you are over the limit; build your client to wait and retry. The OpenRouter 429 guide, Gemini 429 guide and Groq guide cover each one.
Combining free tiers
Because most of these APIs accept the OpenAI chat format, you can switch between them by changing a base URL and key. A router can do it per request: SentioGate automatically falls back to another model if one fails, and adding your own provider keys (including free-tier keys) in the Vault switches you to BYOK at 0% token markup. The BYOK guide explains how that works, and the LLM router hub covers routers in general.
FAQ
Which LLM APIs are free in 2026?
As of 31 Oct 2026, the Google Gemini API Free tier, Groq's Free plan, OpenRouter's :free models, Mistral's Free mode and the SentioGate free community pool. Cerebras offers trial credits instead of a free tier.
Which free LLM APIs need no credit card?
The Gemini API Free tier, Groq's Free plan, OpenRouter's :free models (at the 50-requests-per-day level) and SentioGate's free community pool all start without a card. The Cerebras trial needs a verified payment method.
How do I get a free LLM API key?
Sign up with the provider and create a key in its console: Google AI Studio for Gemini, the GroqCloud console for Groq, or your OpenRouter account for :free models. For SentioGate, create a gateway key in the Vault and send model "auto".
Is GitHub Models still free?
No. GitHub's documentation says GitHub Models was retired as of July 30, 2026.
Does Hugging Face have a free inference API?
Hugging Face Inference Providers gives free accounts no monthly credits, so you must buy credits. PRO accounts get $2 of credits a month.
What are the OpenRouter free model limits?
Free model variants allow 20 requests per minute. The daily cap is 50 requests if you have bought under 10 credits in total, and 1,000 once you have bought 10 or more.
Sources
All accessed 31 Oct 2026.