Free AI APIs
-
Cohere Free API: Embedding and Rerank for RAG
Cohere is the most underrated free AI API for RAG developers. A single free trial key — no credit card, no expiry — unlocks Embed v3 (top-of-MTEB...
-
Groq vs Cerebras vs Gemini Free API: Limits and Speed, Tested
Groq, Cerebras and Gemini re-checked on 2026-09-29: only Groq is free with no card and no expiry; Cerebras is a $5, 30-day trial; Gemini is free but unpublished. Working model IDs and US speeds.
-
Cerebras API: Fast, But the Free Tier Now Needs a Card
Cerebras no longer has a no-card free tier: new accounts get $5 of credit expiring in 30 days, and must add a payment method. Groq measured 527 tok/s and still needs no card.
-
Mistral Free API Tested: 167 tok/s, No Credit Card
We benchmarked Mistral free tier in August 2026: 0.43s to first token and 167 tok/s, second only to Groq. OpenAI-SDK compatible, no credit card.
-
GitHub Models API Is Retired: What to Use Instead
GitHub Models and its API were retired on July 30, 2026, so no API key or token will bring them back. Dates from GitHub's changelog, what the endpoint returns now, and free replacements.
-
Cloudflare Workers AI Free Tier: 10,000 Neurons a Day, Tested
The free Workers plan gives 10,000 neurons a day across about 70 models. Measured on 2026-10-11: a 580-token Llama 3.1 8B reply cost 20 neurons (~490 a day) and one FLUX-1 schnell image cost 173 neurons (~57 a day).
-
Best Free AI APIs: Which Ones Still Work, Tested
Every free LLM API we track, including Groq, Gemini, Mistral, OpenRouter and NVIDIA NIM: free allowance, card and phone rules, and the model IDs that answered, each with the date we checked.
-
OpenRouter Free Tier Tested: 16 Free Models, One API Key
We called all 16 OpenRouter :free models on Sept 29, 2026: 8 answered, 4 hit upstream rate limits, 2 need an agentic harness. Plus a same-model speed test vs NVIDIA direct.
-
DeepSeek API Is No Longer Free: We Checked the Balance
Guides still promise 5M free DeepSeek tokens. We checked a real account in August 2026: granted_balance 0.00 and 402 Insufficient Balance. Here is where to get DeepSeek models free instead.
-
Groq Free API Tested: Speed, Rate Limits and What You Can Build
We benchmarked every free Groq model in August 2026 and re-ran it on 2026-10-11: gpt-oss-120b between 261 and 525 tok/s, real rate limits from the API headers, and what you can actually build on the free tier.