Claude Opus 5.5

Claude Opus 5.5 (anthropic/claude-opus-5.5) — Anthropic’s senior model for the hardest tasks: long code, multi-step agents, fine-grained document analysis. Accepts images and files, always reasons (depth is controllable), has native web search and a generous cache: read $0.25, write $6.25 per 1M tokens.

Parameters and capabilities

Parameter	Value
model	anthropic/claude-opus-5.5 (short alias claude-opus-5.5)
reasoning_effort	low, medium, high, xhigh, max
Disabling reasoning	no — the model always thinks
Images on input	yes (image_url, data URLs)
Files on input	yes (PDF/DOCX/TXT etc., up to 5)
Web search	native (web_search: "native"), sources in annotations
Cache	read $0.25, write $6.25 per 1M

Example

curl https://api.mixen.ai/v1/chat/completions \
  -H "Authorization: Bearer $MIXEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-opus-5.5",
    "reasoning_effort": "high",
    "messages": [{"role": "user", "content": "Refactor this module keeping the public API: <code>"}]
  }'

Pricing

Per 1M tokens: input $5, output $25, cache read $0.25, cache write $6.25. Current prices — catalog.

Strengths and limitations

• Reasoning is always on — its tokens are billed on every request; budget for them.
• Reasoning eats max_tokens: set a generous ceiling for long answers.
• The context cache fires automatically on a repeated prefix; keep the system prompt stable.
• The same model serves the Anthropic protocol (/v1/messages, Claude Code).

All models — in the catalog. The general workflow — chat guide.