Anthropic · 2026-09-22 · seismic
Claude Opus 5.5 — Anthropic's new flagship costs 40% less than Opus 5
Claude Opus 5.5 is Anthropic's new flagship model, priced at $4 per million input tokens and $20 per million output tokens — 40% less to run than Opus 5. It scores 66.4% on Terminal-Bench 4.0 with a 1M-token context window.
Anthropic's new top model matches Fable 5.1 on most work while costing 40% less per token than Opus 5.
Quick facts
| Maker | Anthropic |
|---|---|
| Model ID | claude-opus-5-5 |
| Context window | 1M tokens |
| Max output | 128K tokens |
| Knowledge cutoff | Jun 2026 |
| Thinking | Adaptive, always on |
| Availability | Claude API, Amazon Bedrock, Google Cloud, Microsoft Foundry |
Benchmarks
| Claude Opus 5.5 | 66.4% | |
|---|---|---|
| GPT-6 Astra | 57.9% | |
| Claude Fable 5.1 | 55.8% | |
| Claude Opus 5 | 52.3% | |
| GPT-5.6 Sol | 37.3% |
| Claude Opus 5.5 | 54.4% | |
|---|---|---|
| GPT-6 Astra | 53.3% | |
| Claude Fable 5.1 | 50.3% | |
| Claude Opus 5 | 48% | |
| GPT-5.6 Sol | 47.5% |
| Claude Opus 5.5 | 67.7% | |
|---|---|---|
| Claude Fable 5.1 | 65.6% | |
| Claude Opus 5 | 63.6% | |
| GPT-6 Astra | 57.2% |
Pricing
| Input | $4.00 / 1M tokens |
|---|---|
| Output | $20.00 / 1M tokens |
| Cache read | $0.20 / 1M tokens |
| Cache write (5m) | $5.00 / 1M tokens |
| Batch API · input and output | 50% off |
What is it?
Claude Opus 5.5 opens the Claude 5.5 family and takes the top slot for long-running agentic coding and knowledge work. Anthropic prices it at $4 per million input tokens and $20 per million output tokens, against $10 / $50 for Claude Fable 5.1. The model carries a 1M-token context window, a 128K-token output limit, and a June 2026 knowledge cutoff.
How does it work?
Adaptive thinking is always on and cannot be switched off — depth is steered by an `effort` parameter that defaults to `medium` rather than a manual token budget. That change comes with four breaking differences from Opus 5: thinking cannot be disabled, forced tool use returns an error, thinking blocks are pinned to the model and conversation that produced them, and the older `computer_20251124` computer-use tool is rejected on the Claude API and Google Cloud. Text between tool calls now arrives inside `thinking` blocks that are empty at the default `display` setting.
Why does it matter?
Frontier-level results at $4 / $20 change the arithmetic for teams running agents for hours at a time, where token spend — not capability — is usually the limit. Anthropic reports output generated more than 30% faster and fewer tokens used per task, and says one early tester finished a 680,000-line code migration in under a day. Anthropic also reports its highest alignment audit scores for the model, describing it as less likely than recent models to take hard-to-reverse actions and more resistant to prompt injection.
Who is it for?
teams building coding agents and long-running automation
Frequently asked questions
- How much does Claude Opus 5.5 cost?
- Claude Opus 5.5 costs $4 per million input tokens and $20 per million output tokens on the Claude API. Cache reads are $0.20 per million and a 5-minute cache write is $5 per million. Batch API requests take 50% off both sides. That is 40% less to run than Claude Opus 5, and well under Claude Fable 5.1 at $10 / $50.
- What breaks when upgrading from Claude Opus 5 to Claude Opus 5.5?
- Four changes break existing Claude Opus 5 code. Thinking can no longer be disabled, forced tool use returns an error, thinking blocks are tied to the model and conversation that produced them, and the `computer_20251124` computer-use tool is rejected on the Claude API and Google Cloud. Anthropic publishes a migration guide covering each one.
- How does Claude Opus 5.5 compare to Claude Opus 5?
- Claude Opus 5.5 beats Claude Opus 5 across Anthropic's published evaluations: 66.4% versus 52.3% on Terminal-Bench 4.0, 54.4% versus 48.0% on FrontierCode v1.1, and 67.7% versus 63.6% on Humanity's Last Exam. Anthropic also reports output generated over 30% faster and fewer tokens spent per task, at 40% lower cost.
- Where can I run Claude Opus 5.5?
- Claude Opus 5.5 runs on the Claude API, Amazon Bedrock, Google Cloud and Microsoft Foundry, plus Claude Platform on AWS. The model ID is `claude-opus-5-5` everywhere except Amazon Bedrock, which uses `anthropic.claude-opus-5-5`. It also powers claude.ai and the Claude apps, including Claude Code.
Try it
Model ID `claude-opus-5-5` on the Claude API, Amazon Bedrock, Google Cloud and Microsoft Foundry