Models
Routed through the prix proxy. Auth, billing, retries, and streaming handled once — point any agent at api.prix.dev and pick a model by ID.
# List every model the proxy routes
rush http GET /api/v1/models
# Call any of them — same OpenAI-compatible payload
rush http POST /api/v1/chat/completions -d '{
"model": "anthropic/claude-sonnet-4.5",
"messages": [{"role": "user", "content": "ship it"}],
"stream": true
}'Claude family. Long-context reasoning, tool use, vision. Routed via OpenRouter.
anthropic/claude-opus-4.5via openrouteranthropic/claude-sonnet-4.5via openrouteranthropic/claude-sonnet-4via openrouteranthropic/claude-haiku-4.5via openrouterGPT family. Routed directly, not through a meta-router.
openai/gpt-5-codexvia openaiKimi K2 line. Long context, reasoning traces, cheap inference. Routed via OpenRouter.
moonshotai/kimi-k2.5via openroutermoonshotai/kimi-k2-thinkingvia openroutermoonshotai/kimi-k2-thinking-turbovia openrouterM2 series. Reasoning-first models at low per-token cost. Routed via OpenRouter.
minimax/minimax-m2.5via openrouterLocal-only models. Zero per-token cost, billed by compute time on your own host.
ollama/ministral-3via ollamaollama/qwen3.6:35bvia ollamaThe proxy also routes Replicate image models (Nano Banana, Flux Kontext) and video models (Kling v2.6) under the same auth and billing flow.
Read the quickstart