Cloudflare Workers AI available models
The Cloudflare Workers AI models routable through this free, OpenAI-compatible gateway, with the exact ids you can pin. Cloudflare Workers AI runs a permanent free tier.
| Model id | Pin it as |
|---|---|
| @cf/meta/llama-3.3-70b-instruct-fp8-fast | cloudflare/@cf/meta/llama-3.3-70b-instruct-fp8-fast |
| @cf/ibm-granite/granite-4.0-h-micro | cloudflare/@cf/ibm-granite/granite-4.0-h-micro |
| @cf/nvidia/nemotron-3-120b-a12b | cloudflare/@cf/nvidia/nemotron-3-120b-a12b |
| @cf/openai/gpt-oss-120b | cloudflare/@cf/openai/gpt-oss-120b |
The first id listed is what "model": "auto" reaches for on this provider. Availability changes, so read the live list with GET /api/v1/models before hardcoding.
Calling Cloudflare Workers AI
curl https://dreamprompting.com/api/v1/chat/completions \
-H "Authorization: Bearer YOUR_FREE_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "cloudflare/@cf/meta/llama-3.3-70b-instruct-fp8-fast",
"messages": [{"role": "user", "content": "Hello."}]
}'
Get the key on the account page with a Google sign-in and no credit card. Drop the provider prefix and send "model": "auto" to let the gateway pick a healthy provider instead, which is the better default when you care more about the answer arriving than about which company served it. Full parameters are in the API reference.
Other providers
Looking for the wider landscape rather than what is callable here? See all AI models for the full 279-model index by lab.