Claude Opus 5.5

Claude Opus 5.5 is a Claude model for long-running agentic coding and knowledge work. It is part of the claude-opus line, released 2026-09-22, succeeding Claude Opus 5. It has a 1,000,000 token context window and a max output of 128,000 tokens. Weights are not open. It supports reasoning mode and tool calling, and accepts image, pdf, and text inputs, returning text. Price is from $4.00 per million input tokens and $20.00 per million output tokens, with cached input from $0.20 per million tokens. Knowledge cutoff is 2026-06, and it is served by 15 hosts.

Against Claude Opus 5: input $5.00 to $4.00, output $25.00 to $20.00, cached input $0.50 to $0.20.

Range across the hosts that serve it: up to $9.60 per 1M input tokens.

• ·Long context window of 1,000,000 tokens
• ·Max output of 128,000 tokens
• ·Supports reasoning mode and tool calling
• ·Accepts image, pdf, and text inputs
• ·Cached input pricing at $0.20 per million tokens

• Reach for it for long-running agentic coding tasks
• Reach for it for knowledge work with large documents
• Reach for it for tasks needing tool calling and reasoning
• Reach for it for processing images, PDFs, and text

15 hosts serve this model at the price above · the maker's documentation

• 8Getting the most out of Opus 5.5 in Claude and Claude Code2026-10-04
• 8GPT-6.1 Sol (Max) Enters Agent Arena at Rank 5, Approaching Top Models at Lower Cost2026-10-03
• 8Argon aims to return Google to the frontier2026-10-02
• 8[AINews] Opus 5.5 is good at explainer videos2026-09-29
• 8Show HN: Has Anthropic been nerfing their models without disclosure2026-09-28
• 8Claude Opus 5.5 (High) tops Arena Text Arena leaderboard with 1509 points2026-09-27
• 8Kākāpō Party2026-09-27
• 8[AINews] The Future of Latent Space2026-09-26
• 9Back to Claude2026-09-25
• 8Anthropic launches Claude Opus 5.5: optimizing cost for longer, more context-heavy coding sessions2026-09-25

19 signals of the archive mention this model, each scored against the published bar, the 10 most recent shown. This list is rebuilt from the archive at every build, so it cannot go stale.

The registry keeps the six most recent versions of a line. Older ones stay in the file and can be surfaced without a rebuild.

What is the context window size?

The context window is 1,000,000 tokens.

Is the model open weights?

No, it is not open weights.

What is the price per million tokens?

Input from $4.00, output from $20.00, cached input from $0.20 per million tokens.