Free AI API Keys & Official Free Tiers by Developer

Official free API tiers and one-time trial credits offered directly by LLM model developers (Google, Mistral, Anthropic and others) — entered by hand from each developer's own documentation, with a source link and a check date on every entry. Nothing here has been tested with a real API call, and third-party gateways or resellers (OpenRouter, Together, etc.) are out of scope for this page.

How this page is built. Every entry below is read and entered by a person from the developer's own pricing or docs pages — nothing is scraped, inferred, or copied from a third-party list. Each entry links its source(s) and the date someone last checked it; an entry not re-checked in over 90 days is flagged so you can confirm it yourself before relying on it. For some developers we only link to their official page instead of summarizing it — see “Other developers” below.

Anthropic — Claude API new-user credits

Confirmed from official docs (not tested by a real API call) One-time trial credits

Anthropic's pricing FAQ answers "Are there free tiers or trials?" with a small amount of free credits for new users to test the API.

"New users receive a small amount of free credits to test the API." The amount is not stated. (Claude API pricing — FAQ: Are there free tiers or trials?)
Extended trials for enterprise evaluation are available by contacting Anthropic sales. (Claude API pricing — FAQ: Are there free tiers or trials?)

Checked 2026-09-23.

Sources

Claude API pricing — FAQ: Are there free tiers or trials?
Anthropic, Claude API pricing documentation. Quoted in part.

See Anthropic's full model pricing →

Google — Gemini API free tier (Google AI Studio)

Confirmed from official docs (not tested by a real API call) Ongoing free tier

The Gemini API has a Free tier that offers free input and output tokens on a subset of models, with limited access to certain models and rate limits that are not guaranteed.

Free tier includes "Free input & output tokens" and "Limited access to certain models". (Gemini Developer API pricing)
Models listed as "Free of charge" for both input and output on the free tier include gemini-3.8-flash, gemini-3.7-flash, gemini-3.6-flash, gemini-3.5-flash, gemini-3.5-flash-lite, gemini-3.1-flash-lite, gemini-3-flash-preview, gemini-2.5-pro, gemini-2.5-flash, gemini-2.5-flash-lite, Gemma 4 and gemini-embedding-2, plus several TTS, Live and transcription models. (Gemini Developer API pricing)
Some models are "Not available" on the free tier, including Gemini 3.1 Pro, gemini-3.1-pro-preview-customtools and the Gemini Omni Flash and image models. (Gemini Developer API pricing)
On the free tier, content is "used to improve our products". (Gemini Developer API pricing)
The Free usage tier applies to an "Active project or free trial"; paid tiers require setting up billing. (Gemini API rate limits)
Per-model free-tier rate limits are not listed on the page: "View your active rate limits in AI Studio". "Specified rate limits are not guaranteed and actual capacity may vary." (Gemini API rate limits)

Checked 2026-09-23.

Sources

Gemini Developer API pricing — page last updated 2026-09-23 — CC BY 4.0
Google, "Gemini Developer API pricing", ai.google.dev, CC BY 4.0. Summarized and quoted in part; model list shortened.
Gemini API rate limits — page last updated 2026-09-02 — CC BY 4.0
Google, "Rate limits", ai.google.dev, CC BY 4.0. Summarized and quoted in part.

See Google's full model pricing →

Mistral — Mistral Studio Free mode

Confirmed from official docs (not tested by a real API call) Ongoing free tier

Mistral Studio enables API access in Free mode by default, with no credit card required; usage and rate limits apply.

"Free mode: API access is enabled by default with no credit card required. Usage and rate limits apply." (Activate Studio and generate an API key (Mistral docs source))
Rate limits are set per Workspace (requests per second, tokens per minute, tokens per month); your current limits are shown in the Admin Panel under API > Limits. (Manage Workspace usage and limits (Mistral docs source))

Checked 2026-09-23.

Sources

Activate Studio and generate an API key (Mistral docs source) — page last updated 2026-09-22 — Apache-2.0
Mistral AI, platform-docs-public (Apache-2.0). Quoted in part.
Manage Workspace usage and limits (Mistral docs source) — page last updated 2026-09-22 — Apache-2.0
Mistral AI, platform-docs-public (Apache-2.0). Summarized.

See Mistral's full model pricing →

Other developers

We don't summarize these developers' free-tier or trial terms here. Check their official pages directly.

Free tier or trial terms: see Alibaba's official pricing page.
Free tier or trial terms: see DeepSeek's official pricing page.
Free tier or trial terms: see OpenAI's official pricing page.
Free tier or trial terms: see Z AI's official pricing page.
Free tier or trial terms: see xAI's official pricing page.

Methodology: entries are entered by hand from each developer's own documentation and not verified by a real API call; each entry shows the date it was last checked. Free-tier terms change without notice — always confirm current limits on the developer's own page before relying on a free tier in production. If something below looks out of date, the linked official page is the source of truth.