Claude Opus 5.5 API: Anthropic's new flagship, priced below Opus 5

Claude Opus 5.5 is Anthropic's newest flagship, and it is priced below the model it replaces. It went live on FastMetal on September 22, 2026, as anthropic-claude-opus-5-5.

New generations have usually raised per-token rates. Opus 5.5 lowers both the input and the output rate compared with Claude Opus 5. It is billed per million input and output tokens; current rates are on the Claude Opus 5.5 and Claude Opus 5 pages.

What it supports

• A 1M-token context and up to 128K output tokens.
• Text and images in the same request.
• Tool calling (function calling) and structured output with a JSON schema.
• A settable reasoning effort: the FastMetal gateway accepts the reasoning object.
• No training on your prompts.

Calling it

It is OpenAI-compatible, so existing code only needs a new model value:

curl https://api.fastmetal.ai/v1/chat/completions \
  -H "Authorization: Bearer $FASTMETAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic-claude-opus-5-5",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Should you switch from Opus 5?

If you already call Opus 5, changing model to anthropic-claude-opus-5-5 makes the same work cheaper, with the same context length and the same output limit. Move a share of your requests first and compare the outputs before you switch everything. For how Opus 5 itself held up, see our Claude Opus 5 review.

Opus models sit at the top of the price range. Billing is prepaid: you load a balance by card, from $5, and you can never spend more than you loaded. See pricing, and everything else we serve in the model catalog.