Skip to main content
Radar

Every change to the models you run, with its source and its date. Releases, price changes, retirements, API changes, and incidents.Every change to the models you run, with its source.

Follow model changes
Free tiers and entry limits

Free LLM API tiers, with the limits that apply

Four providers, four different answers. Google's Gemini API has a real free tier without billing, and its per-model limits now live in the AI Studio console. OpenRouter's free models are metered per request. DeepSeek and Anthropic have no free tier: DeepSeek is prepaid, Anthropic gives an unstated signup credit. Every limit below carries the provider source and the date we read it.

Limits read from provider pages

By provider

OpenRouter

Free tier

Models whose ID ends in :free cost nothing per token and are limited per account: 20 requests per minute, and 50 requests per day until the account has bought $10 of credits in its lifetime, after which the daily limit is 1,000. Read 2026-09-11.

Limits and sources

OpenAI

Free tier

OpenAI's rate-limits guide lists a Free usage tier for organisations that have not paid, available only in an allowed geography, with a $100 a month usage limit. Read 2026-09-14.

Limits and sources

Mistral AI

Free tier

Mistral's help article describes Free mode as the default with limited limits for testing and prototyping, and says the current limits, usage and tier are shown on the Limits page of the Admin panel (the console). Read 2026-09-14.

Limits and sources

Groq

Free tier

Groq publishes sample Free-plan limits per model, applied at the organisation level, and says the current exact limits for an organisation are on the limits page in account settings, with possible exceptions. Read 2026-09-14.

Limits and sources

Together AI

No free tier

Together's billing page states that Together AI does not currently offer free trials, and that access to the platform requires a minimum $5 credit purchase. Read 2026-09-14.

Limits and sources

Cerebras

Signup credit

New Cerebras accounts receive $5 in free credits after adding a verified payment method; the credits expire 30 days after they are granted, and Playground and API access stay inactive until a payment method is added. Read 2026-09-14.

Limits and sources

xAI

No free tier

xAI's API page states that API usage is billed per token rather than free, and that the Playground comes with every Console account. Read 2026-09-14.

Limits and sources

How these pages are kept honest

Every limit is copied from the provider's own page on the date shown beside it, with a link to that page. When a provider stops publishing a number, as Google did for the free tier's per-model limits, the row says so and stays unpublished; we do not fill it from a forum post or another tracker.

A free tier is a provider change. Rows are re-read before a page is published, and the paid rates that follow each tier are on the provider pricing hubs and the LLM pricing comparison.

Questions

Which LLM APIs have a free tier?

Of the eleven providers on this page, Google’s Gemini API, Groq, Cohere’s Trial key and Mistral’s Free mode are free tiers without billing, OpenRouter offers free model variants metered by request, and OpenAI lists a Free usage tier by geography. Cerebras grants a $5 credit after a card is added. DeepSeek, Anthropic, Together AI and xAI publish no free tier.

Is there a free LLM API for production use?

Not one you can plan a workload on. Free tiers are rate-limited per minute or per day, Google now shows its free limits only in the AI Studio console, and every provider can change them without notice. Treat a free tier as a prototyping budget and price the production workload on the paid rates.

How often do free tiers change?

Often enough that the date matters. Google moved its Pro models to paid in April 2026 and stopped publishing free-tier limits on its docs page, and OpenRouter ties its daily cap to purchase history. Every limit here shows the date we read it from the provider’s page, and rows are re-read before they are published.