10,000 Neurons/day free
Catch: Requests fail with an error past the daily cap (resets 00:00 UTC); some models need a paid billing method.
| # | Tool / free plan | Limit | Card | Checked | Compare |
|---|---|---|---|---|---|
| 022 | Cloudflare Workers AI AI APIs 10,000 Neurons/day free Catch: Requests fail with an error past the daily cap (resets 00:00 UTC); some models need a paid billing method. |
Daily neuron cap | 2026-10-13 | ||
| 023 | Cohere AI APIs Trial API key for evaluation Catch: Trial keys are rate limited and may not be used for production or commercial purposes. |
Rate-limited, non-production use | 2026-10-13 | ||
| 046 | Google AI Studio AI APIs Free tier for Gemini API in AI Studio Catch: Free-tier content is used to improve Google products; newest models such as Gemini 4 Argon are not available. |
Rate limits per model; data may be used to improve products | 2026-10-13 | ||
| 052 | Groq AI APIs Free API plan for open models such as gpt-oss-120b, Qwen and Whisper Catch: Low per-model rate limits. |
Per-model rate limits, e.g. 30 requests/min and 1,000 requests/day on gpt-oss-120b | 2026-10-13 | ||
| 057 | Hugging Face AI APIs Hub hosting, Spaces on free CPU, limited inference credits Catch: Not stated by vendor |
Small monthly inference credits | 2026-10-13 | ||
| 071 | LM Studio AI APIs Free desktop app to run local LLMs Catch: Not stated by vendor |
Local hardware limits | 2026-10-13 | ||
| 082 | Mistral La Plateforme AI APIs Free experiment tier for API Catch: Not stated by vendor |
Rate limits | 2026-10-13 | ||
| 091 | NVIDIA NIM AI APIs Free serverless API endpoints for open models (e.g. Nemotron, DeepSeek, Kimi) for prototyping Catch: Not for production use; production requires an NVIDIA AI Enterprise license (from $4,500 per GPU per year). |
Development, testing and research only; production needs an NVIDIA AI Enterprise license | 2026-10-14 | ||
| 093 | Ollama AI APIs Free local model runner Catch: Starter models only; more cloud models need added credits. |
Local hardware limits; cloud paid | 2026-10-13 | ||
| 095 | OpenRouter AI APIs Chat and API access to 25+ free models from 4 free providers Catch: No budgets, spend controls or prompt caching on Free. |
No prompt caching, budgets or management keys on Free | 2026-10-13 |
Cohere#023
Trial API key for evaluation
Catch: Trial keys are rate limited and may not be used for production or commercial purposes.
Google AI Studio#046
Free tier for Gemini API in AI Studio
Catch: Free-tier content is used to improve Google products; newest models such as Gemini 4 Argon are not available.
Groq#052
Free API plan for open models such as gpt-oss-120b, Qwen and Whisper
Catch: Low per-model rate limits.
Hugging Face#057
Hub hosting, Spaces on free CPU, limited inference credits
Catch: Not stated by vendor
LM Studio#071
Free desktop app to run local LLMs
Catch: Not stated by vendor
Free experiment tier for API
Catch: Not stated by vendor
NVIDIA NIM#091
Free serverless API endpoints for open models (e.g. Nemotron, DeepSeek, Kimi) for prototyping
Catch: Not for production use; production requires an NVIDIA AI Enterprise license (from $4,500 per GPU per year).
Ollama#093
Free local model runner
Catch: Starter models only; more cloud models need added credits.
OpenRouter#095
Chat and API access to 25+ free models from 4 free providers
Catch: No budgets, spend controls or prompt caching on Free.