Claude Opus 5.5
Claude Opus 5.5 by Anthropic: frontier foundation model performing at Claude Fable 5.1 level, featuring 1M context, 40% cost reduction, and SOTA agentic coding scores.
Model Overview
Capabilities, design details, and architectural traits
Claude Opus 5.5 - Frontier Model for Complex Agentic Coding and Knowledge Work
Claude Opus 5.5 is Anthropic's flagship model in the Claude 5.5 generation, introduced on September 22, 2026. It delivers intelligence on par with Claude Fable 5.1 across most workloads while costing 40% less to run than Opus 5, with output token generation speed increased by over 30%.
| Trait | Detail |
|---|---|
| Context Window | 1,000,000 tokens (1M tokens) native capacity |
| Agentic Coding | SOTA performance: 66.4% on Terminal-Bench 4.0, 54.4% on FrontierCode v1.1, and 57.8% on CursorBench 4.0 |
| Cost & Efficiency | 40% cheaper overall ($4/M input, $20/M output, and $0.20/M cache reads - a 60% cache reduction) |
| Computer Use & Vision | 81.8% on OSWorld 2.0 and 89.0% on Chartography visual recognition |
| Safety & Verification | Highest score to date on Anthropic's Automated Behavioral Audit; integrates Cyber and Life Sciences verification safeguards |
Frontier Performance & Real-World Agentic Engineering
Opus 5.5 introduces major architectural refinements in multi-hour autonomous execution, software refactoring, and code migration. In real-world enterprise evaluations, it completed a 680,000-line code migration in less than a day and achieved a 39/40 success rate optimizing full-stack web application performance. It communicates with clearer, more natural phrasing and supports adaptive thinking up to maximum effort for complex multidisciplinary challenges.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Anthropic | Anthropic | Meta |
| Release Date | September 22, 2026 | September 1, 2026 | September 2, 2026 |
| Knowledge Cutoff | - | - | - |
| Context & Limits | |||
| Context Window | 1M | 1M | 1M |
| Pricing (per 1M tokens) | |||
| Input Pricing | $4 | $10 | $1.25 Best Input Pricing |
| Output Pricing | $20 | $50 | $4.25 Best Output Pricing |
| Modalities | |||
| Inputs | textimagefile | textimagefile | textimagefilevideo |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 57.6 Best Intelligence Index | 53.4 | 53.0 |
| Coding Index | - | 81.6 Best Coding Index | 76.3 |
| Agentic Index | - | 57.9 Best Agentic Index | 55.6 |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.