Model Details
Claude Sonnet 5.5
Quick answer: Claude Sonnet 5.5 is the second model in Anthropic's Claude 5.5 family, released September 28, 2026. It's a faster, lower-cost complement to Claude Opus 5.5 — Anthropic says it runs 30%+ faster and costs up to 30% less per task than Claude Sonnet 5, at the same $2/$10 per Mtok list price.
At a Glance
- Where it leads: Terminal-Bench 4.0 (70.6%, up from Sonnet 5's 10.3%), scores within ~2 points of Claude Opus 5.5 on FrontierCode v1.1 at Xhigh effort.
- Where it lags: Stays below Claude Opus 5.5 on GDPVal-AA v2.1, Humanity's Last Exam with tools, and OSWorld 2.1 — Anthropic positions Opus 5.5 as the stronger model for complex, open-ended work.
What Claude Sonnet 5.5 Is
Claude Sonnet 5.5 is Anthropic's faster, cheaper everyday-work model, sitting below Claude Opus 5.5 in the Claude 5.5 family. It targets well-scoped tasks — bug fixes, polished documents/slides/spreadsheets, fast iteration — rather than the complex, sustained-judgment work Opus 5.5 is built for. Claude Haiku 5.5 is expected to complete the family in the following weeks.
Specifications
| Organization | Anthropic |
| License | Proprietary (API only) |
| Release date | September 28, 2026 |
| Modality | Text |
| Knowledge cutoff | Not disclosed |
Public Benchmark Scores
| Benchmark | Score | Source | Date |
|---|---|---|---|
| Terminal-Bench 4.0 | 70.6% | Anthropic | Sep 2026 |
| FrontierCode v1.1 (Main) | 52.1% (xhigh effort) | Anthropic | Sep 2026 |
| GDPVal-AA v2 | 1,844 Elo (v2.1-anchored) | Anthropic | Sep 2026 |
| AA-Briefcase v1.1 | 1,811 | Anthropic | Sep 2026 |
| Humanity's Last Exam | 64.5% (with tools) | Anthropic | Sep 2026 |
| OSWorld 2.0 | 80.1% (partial) | Anthropic | Sep 2026 |
| CursorBench 4.0 | 55.5% | Anthropic | Sep 2026 |
| Chartography | 61.6% (no tools) | Anthropic | Sep 2026 |
Scores as reported in Anthropic's own Claude Sonnet 5.5 announcement (anthropic.com/claude-sonnet-5-5, September 2026); not independently verified by Benchgen.