Claude Sonnet 5.5
Anthropic · Released 28 Sep 2026 · claude-sonnet-5-5
$2.00
Input price / 1M tokens
#36 of 61, cheapest first
$10.00
Output price / 1M tokens
#39 of 61, cheapest first
1M
Context window
#29 of 71
139
Output tokens / second
#20 of 61
70.6%
Terminal-Bench 4.0
#1 of 10
46.2%
FrontierCode 1.1
#4 of 6
64.5%
Humanity’s Last Exam (with tools)
#3 of 16
Summary
Claude Sonnet 5.5 is a proprietary model from Anthropic, released on 28 Sep 2026.
At $2.00 input and $10.00 output per million tokens, it is #36 of 61 on input price, cheapest first.
Its 1M-token context window ranks #29 of 71.
Measured output speed is 139 tokens per second, #20 of 61.
Anthropic's current Sonnet model, described as the best combination of speed and intelligence and a faster, lower-cost complement to Claude Opus 5.5 for well-scoped everyday tasks.
Benchmarks · 8 reported
| Benchmark | Score | Bar | Rank | Source |
|---|---|---|---|---|
| Terminal-Bench 4.0agentic coding; effort not stated | 70.6% | #1 of 10 | Anthropic Sonnet 5.5 launch postSelf-reported | |
| FrontierCode 1.1max effort | 46.2% | #4 of 6 | Anthropic Sonnet 5.5 launch postSelf-reported | |
| CursorBench 4.0agentic coding | 55.5% | – | Anthropic Sonnet 5.5 launch postSelf-reported | |
| GDPval-AA v2.1knowledge work | 1,844 | Elo-style | – | Anthropic Sonnet 5.5 launch postSelf-reported |
| AA-Briefcase v1.1knowledge work | 1,811 | Elo-style | – | Anthropic Sonnet 5.5 launch postSelf-reported |
| Humanity’s Last Exam (with tools)with tools | 64.5% | #3 of 16 | Anthropic Sonnet 5.5 launch postSelf-reported | |
| OSWorld 2.1partial-credit score | 80.1% | – | Anthropic Sonnet 5.5 launch postSelf-reported | |
| Chartographyno tools | 61.6% | – | Anthropic Sonnet 5.5 launch postSelf-reported |
Self-reported means the lab ran the test itself. A rank appears only where several labs report the same version of a benchmark.
Similar models
Models with the most benchmarks in common with Claude Sonnet 5.5, and the closest scores.
Terminal-Bench 4.0
#1 of 10Long jobs in a real terminal
- Claude Sonnet 5.570.6
- Claude Opus 5.566.4
- Claude Fable 5.155.8
- Claude Opus 552.3
- Claude Sonnet 510.3
- GPT-6 Astra57.9
FrontierCode 1.1
#4 of 6Patches good enough to merge
Humanity’s Last Exam (with tools)
#3 of 16Expert questions, with search and code
- Claude Sonnet 5.564.5
- Claude Opus 5.567.7
- Claude Fable 5.165
- Claude Opus 563.6
- Claude Sonnet 554.9
- GPT-6 Astra57.2
Compare Claude Sonnet 5.5 side by side
Pre-filled with the two closest models. Swap any of them.
Sources
- price, cache prices, batch, tokenizer: platform.claude.com/docs/en/about-claude/pricing, checked 2 Oct 2026
- context, max output, release, api id, knowledge cutoff, modalities, min cacheable prompt: platform.claude.com/docs/en/models/sonnet-5-5/overview, checked 2 Oct 2026
- status, retirement commitment (not sooner than 2027-09-28): platform.claude.com/docs/en/about-claude/model-deprecations, checked 2 Oct 2026
- refusal classifiers and refusal billing: platform.claude.com/docs/en/build-with-claude/refusals-and-fallback, checked 2 Oct 2026
- benchmarks, positioning: anthropic.com/claude-sonnet-5-5, checked 2 Oct 2026
- output speed: artificialanalysis.ai/models/claude-sonnet-5-5, checked 2 Oct 2026
Building on Claude Sonnet 5.5?
Get the architecture right before the bill arrives.
We size caching, routing and fallbacks for your workload on a free call, and tell you where a cheaper model is good enough.