Overview
Claude Sonnet 5.5 was released on September 28, 2026. Anthropic positions it as a faster, lower-cost complement to Claude Opus 5.5 for well-scoped everyday coding and knowledge work. Compared with Sonnet 5, Anthropic reports over 30% faster output and up to 30% lower cost per task in its testing; the latter is not a reduction in the standard input or output token rates. (anthropic.com)
Capabilities and uses
- Supports coding, bug fixing, data analysis, content creation, visual understanding and agentic tool use. Suggested uses include producing documents, presentations and spreadsheets. (platform.claude.com)
- Has a 1 million-token context window and a standard maximum output of 128,000 tokens. The Message Batches API supports up to 300,000 output tokens with the specified beta header. The documented knowledge cutoff is June 2026. (platform.claude.com)
- Anthropic reports a 70.6% score on Terminal-Bench 4.0. This is a vendor-reported evaluation, not a guarantee of performance on a particular task. (anthropic.com)
Limitations and availability
- Opus 5.5 is positioned for more complex work requiring careful judgment. Sonnet 5.5 has safeguards for a narrow set of high-risk cybersecurity and biology requests. (anthropic.com)
- Adaptive thinking is enabled by default. Setting
temperature,top_portop_kreturns an API error; forced tool choices of typeanyortoolare also unsupported. Migration from Sonnet 5 may require code changes. (platform.claude.com) - Available through the Claude API and several cloud platforms. Model identifiers and platform-specific terms can differ; for example, Amazon Bedrock uses
anthropic.claude-sonnet-5-5. (platform.claude.com)
Pricing
Claude API, USD per 1 million tokens; checked November 16, 2026. The following first-party prices apply across the full context window under global routing: (assets.claude.com)
| Billing category | Price |
|---|---|
| Standard input | $2.00 |
| Standard output | $10.00 |
| Prompt-cache write, 5-minute duration | $2.50 |
| Prompt-cache write, 1-hour duration | $4.00 |
| Prompt-cache read | $0.10 |
| Batch API input | $1.00 |
| Batch API output | $5.00 |
The asynchronous Batch API discount and cache pricing can be combined. US-only inference has a 1.1× multiplier on token categories; for example, standard input and output are respectively $2.20 and $11.00 per million tokens. Partner-operated cloud platforms have independent regional pricing, and server-side tools may incur additional charges. (assets.claude.com)