cuna inference
Pick The Model. Pay Per Token.
Serverless. Metered by the token.
- gpt-oss-120bContext 131K · Reasoning · Fn calling~640tok/sinput$0.30/Mcached in—output$1.20/MShow the request ›Hide the request ›gpt-oss-120b-fast · not open yet
curl https://inference.getcuna.com/v1/chat/completions \ -H "Authorization: Bearer $CUNA_API_KEY" \ -H "Content-Type: application/json" \ -d '{"model": "gpt-oss-120b-fast", "messages": [{"role": "user", "content": "Hello"}]}'
- GLM-5.2Context 128K · Coding + reasoning · Z.ai's flagship, served lossless~280tok/sinput$1.00/Mcached in—output$4.00/MShow the request ›Hide the request ›glm-5.2 · not open yet
curl https://inference.getcuna.com/v1/chat/completions \ -H "Authorization: Bearer $CUNA_API_KEY" \ -H "Content-Type: application/json" \ -d '{"model": "glm-5.2", "messages": [{"role": "user", "content": "Hello"}]}'
- Kimi K3Context 1M · Reasoning · Vision · Moonshot's 2.8T hybrid MoE~135tok/sinput$2.50/Mcached in—output$12.50/MShow the request ›Hide the request ›kimi-k3 · not open yet
curl https://inference.getcuna.com/v1/chat/completions \ -H "Authorization: Bearer $CUNA_API_KEY" \ -H "Content-Type: application/json" \ -d '{"model": "kimi-k3", "messages": [{"role": "user", "content": "Hello"}]}'
USD per million · throughput approximate