Every model you can run here, priced per million tokens, and 3.2M more to read about, download or deploy to a GPU. One key, and one conversation that follows you when you switch.
Routes each request to the cheapest model that can handle it — one that sees images if you send one, calls tools if you send tools, and fits your conversation — and moves up to a stronger model if that one fails. No fixed price: you pay for whichever models ran, at their listed prices, capped by max_spend (default $0.10 per request). Without a verify schema it escalates on errors and refusals only, not on answer quality.
Claude Opus 5, served on Amazon Bedrock.
Claude Sonnet 5, served on Amazon Bedrock.
Llama 3.3 70B, served on Amazon Bedrock.
Claude Haiku 5.5, served on Amazon Bedrock.
Claude Opus 5.5, served on Amazon Bedrock.
Claude Sonnet 5.5, served on Amazon Bedrock.
Gemma 3 27B, served on Amazon Bedrock.
Magistral Small 1.2, served on Amazon Bedrock.
Mistral Large 3, served on Amazon Bedrock.
Pixtral Large, served on Amazon Bedrock.
Kimi K2 Thinking, served on Amazon Bedrock.
Every model you can run here, priced per million tokens, and 3.2M more to read about, download or deploy to a GPU. One key, and one conversation that follows you when you switch.
Nothing in the runnable catalogue matches “nvidia/bigvgan_v2_44khz_128band_512x”. The library below may.
Searching the library…