Free AI
Free AI
Free AI services; entries with a provider config import into ChatSpeed
Alibaba Cloud Model Studio Free Quota
VerifiedFirst activation auto-grants per-model new-user free quota (usually 1M tokens per model, valid 90 days, Beijing region, no real-name needed); Qwen Code OAuth adds an independent 2,000 calls/day.
Cerebras Free Tier
VerifiedUltra-fast OpenAI-compatible inference (up to ~2,000 tokens/s) with a free tier of roughly 1M tokens/day per model (e.g. Llama 3.3 70B); no credit card required.
Google Gemini API Free Tier
VerifiedFree tier covers Flash / Flash-Lite family models (e.g. Gemini 2.5 Flash, Gemini 2.5 Flash-Lite, newer 3.x Flash Lite) with free input/output tokens; quotas are relatively generous for daily use and shown in AI Studio.
Groq Free API
VerifiedPermanent free tier, no credit card required: e.g. Llama 3.1 8B gets 30 RPM / 14,400 requests per day; Llama 3.3 70B, Llama 4 Scout, GPT-OSS 120B get 30 RPM / 1,000 RPD; Kimi K2/Qwen3 32B get 60 RPM / 1,000 RPD.
OpenRouter Free Models
VerifiedFree model variants (model IDs with the `:free` suffix) can be called free of charge through an OpenAI-compatible API; new accounts get 20 requests/minute and 50 requests/day, raised to 1,000 requests/day after $10 in lifetime credit purchases.
SenseNova Token Plan (Free Beta)
VerifiedFree open beta: the general-credit pool and the Flash-Lite credit pool each provide a limited-time supply of 60,000 credits per rolling 5 hours and 600,000 credits per rolling week; Flash-Lite usage of dedicated credits earns a 1:1 general-credit rebate (rules effective 2026-08-28).
SiliconFlow Free Models & Credit
VerifiedNew users get about ¥14 (≈20M tokens) credit, and after real-name verification can use many permanently free open models (DeepSeek, Qwen, Llama, GLM etc.); free models cost 0 with fixed rate limits.
Zhipu GLM Flash Free Models
VerifiedGLM-4-Flash is announced as permanently free with basically no call-volume limit (newer GLM-4.7-Flash is also free); sign-up also grants learning credits for paid GLM models (community reports ~20M tokens).
Agnes AI Free API
In reviewOpenAI-compatible multimodal gateway (text/image/video) that grants free API credits on sign-up; several models run free (Agnes 2.0/2.5 Flash, Agnes Image 2.1 Flash), no credit card required.
Cloudflare Workers AI Free Quota
In reviewCloudflare Workers AI grants a daily free allocation (about 10,000 Neurons/day) across ~80 models (Llama, Mistral, Qwen, Gemma, GPT-OSS, FLUX etc.); accounts are free and limits can be adjusted to avoid overage.