Free AI Model APIs
| Model / Platform | Cloudflare Workers AI64 models | NVIDIA API Catalog36 models | OpenCode Zen / Console13 models | Vercel AI Gateway10 models |  OpenRouter6 models |  Respan3 models |
|---|---|---|---|---|---|---|
| — | — | — | ||||
| — | — | — | — | |||
| — | — | — | — | |||
| — | — | — | — | |||
| — | — | — | — | — | ||
| — | — | — | — | — | ||
| — | — | — | — | — | ||
| — | — | — | — | — | ||
38 deals · page 3 of 4
Current offers
Free tier (5M tokens on sign-up, then 1M a month)
Privatemode's encrypted AI free tier gives 5M tokens on sign-up and 1M more each month, no card needed
Privatemode is an end-to-end encrypted AI service with a free tier that needs no credit card. New sign-ups get a one-time 5M-token quota, then 1M input and 1M output tokens free every month. Both the web chat and the API are included, with models such as GLM-5.3, GLM-5.3-Flash, and gpt-oss-120b.
Free
Bilibili opens a free Index-Translate API with no key needed for 150 languages
Bilibili's Index team has opened a free public API for the Index-Translate-35B-A3B translation model. It is OpenAI-compatible and needs no sign-up or API key, and the model covers 150 languages. Point your client's base URL at https://index-translate.bilibili.com/v1 to start. No end date has been announced for the free access.
Free
Laya free on Vercel AI Gateway through October 31
Convai Innovations’ Laya decision model is free to call on Vercel AI Gateway through October 31, 2026. Use model ID convaiinnovations/laya with an AI Gateway API key—built for classification, routing, and guardrail checks.
Free tier: 1,000 credits/month; Hobby from about $16/month
Firecrawl: claim 20,000 free credits with a new-user code
Firecrawl is giving eligible new users 20,000 credits with the code CONVEXALLGAS. Register, open Settings → Billing → Coupons, and redeem the code. The normal free tier includes 1,000 credits per month.
Sol $2/$10 · Luna $0.10/$0.50 per million tokens · about half of GPT-5.6
OpenAI GPT-6 Sol and Luna: permanently half-price API rates plus banked resets
OpenAI's GPT-6 Sol and Luna API models are priced at about half the comparable GPT-5.6 rates. OpenAI also says Plus, Pro, and Business users will receive a banked usage reset; rollout and model access may vary by account.
Other offers & history
Free through Oct 22 · normally $0.15/M input tokens
MiniMax M3.1 Flash, the model behind Space Bunny, is free on Ollama Cloud through Oct 22
Ollama and MiniMax made MiniMax M3.1 Flash free on Ollama Cloud until October 22, 11:59 PM UTC, with $0 input and output. It has a 976K-token context window, image input, tool calling, and thinking. Sign in to Ollama and one command launches it in Claude Code, Codex, or OpenCode.
Free (limited time) · plans from $1/mo
MiniMax M3.1 Flash, the model behind Space Bunny, is free on Command Code from the $1 Go plan up
The free run of MiniMax M3.1 Flash on Command Code ended around October 22; the model is now billed at $0.15 per million input tokens and $0.60 per million output tokens. It keeps its 1M-token context window.
Free through Oct 22
MiniMax M3.1 Flash, the model behind Space Bunny, is free on Vercel AI Gateway through Oct 22
Vercel AI Gateway added MiniMax M3.1 Flash for free as minimax/minimax-m3.1-flash through October 22. Input and output cost $0. It has a 1M-token context window, up to 524K output tokens, tool calling, image input, and reasoning stays on by default.
Free (rate limited)
MiniMax M3.1 Flash, the model behind Space Bunny, is free on OpenRouter with 1M context and multimodal input
OpenRouter added a free MiniMax M3.1 Flash endpoint as minimax/minimax-m3.1-flash:free. Input and output cost $0, with rate limits on the free endpoint. It accepts text, image, and video, has a 1M-token context window, up to 524K output tokens, and reasoning stays on by default.
Free
MiniMax M3.1 Flash, the model behind Space Bunny, is free on OpenCode Zen and Go with 1M context
MiniMax's new M3.1 Flash model, released on October 19, is on OpenCode as MiniMax M3.1 Flash Free. You can pick it in both Zen and Go, and input, output, and cache reads cost nothing. In Go it doesn't use your plan quota and has no request caps. It supports 1M tokens of context, takes image and video input, and lets you tune thinking depth.
Free
Cat Jam stealth coding model is free on Vercel AI Gateway with a 1M-token context window
Cat Jam, a new stealth reasoning model on Vercel AI Gateway, is free to use during its stealth period. Call it as stealth/cat-jam. It is built for coding and agentic tasks, supports function calling and structured outputs, has a 1M-token context window, and works in Claude Code, Codex, Cursor, and other coding agents.
GMI Cloud makes Qwen 3.8 Max, Qwen 3.8 Flash and Wan 3.0 free for a limited time if your balance is at least $10
GMI Cloud and Alibaba made Qwen 3.8 Max, Qwen 3.8 Flash and the Wan 3.0 video model free for developers. On October 7, GMI extended the window by 7 more days, which runs it to around October 14. You need at least $10 in your GMI balance, and that $10 is not spent, so you can use it on other models later.