Claude Opus 5.5 now available on AI Gateway
By Steven Van ·
Anthropic's Opus 5.5 lands on Vercel's AI Gateway at $4/$20 per million tokens, with adaptive thinking and no more forced tool calls.
Claude Opus 5.5 from Anthropic is now available on Vercel's AI Gateway, with the biggest gains over Opus 5 in agentic coding, long-running agent tasks, and knowledge work. Anthropic says it matches Fable 5.1 on performance while running about 30% faster and 40% cheaper than Opus 5 per task, and that it reports back in plain language on what it did, what it found, and what it needs next during long-running work.
Two API changes can turn previously valid requests into HTTP 400 errors:
- Thinking is always adaptive: requests that disable thinking or set a fixed thinking budget are rejected. Use effort and prompting to steer it instead.
- Forced tool use is retired: requests can no longer require a tool call or force a specific tool. Use structured outputs for JSON, and catch and retry misses in your own harness.
Pricing is $4 per million input tokens and $20 per million output tokens. The model has a 1M-token context window, returns up to 128K output tokens, and has a June 2026 knowledge cutoff. It's available as anthropic/claude-opus-5.5 across the AI SDK, the OpenAI-compatible Chat Completions API, the Anthropic Messages API, and coding agents connected to AI Gateway, with a faster anthropic/claude-opus-5.5-fast variant.