Anthropic
Claude Sonnet 5.5
Anthropic's September 28, 2026 model and the second in the Claude 5.5 family after Opus 5.5, built for coding and agent workflows at Sonnet pricing: $2 input and $10 output per million tokens, unchanged from Sonnet 5, with cache reads at $0.10 (cut from $0.20 in October 2026) and cache writes at $2.50. It accepts text and image input, provides a 1M-token context window and up to 128K output, and keeps thinking on with adjustable effort (medium default in Claude apps, high on the Platform). Anthropic reports roughly 30% faster and about 30% cheaper per task than Sonnet 5, Terminal-Bench 4.0 of 70.6% (up from Sonnet 5's 10.3%) and near-Opus-5.5 results on GDPval-AA, and it is the first Sonnet released under Opus-class cyber safeguards. The benchmark figures above are Anthropic's own; Artificial Analysis measurements are shown separately.
- Provider
- Anthropic
- Context
- 1MCatalog · As of Oct 21, 2026Within 30d
- yno subscription tier
- Pro
- Released
- Sep 28, 2026
- Speed
- Fast
- Reasoning
- Expert
- Modality
- Multimodal
Benchmarks
Quality
- AA Intelligence Index
Speed and latency
- Output speed
- 135tokens/sConfiguration: Prompt length: 1,000 · Parallel queries: 1Artificial Analysis · Date unknownAge unknown
- Time to first token
- 282.51sConfiguration: Prompt length: 1,000 · Parallel queries: 1Artificial Analysis · Date unknownAge unknown
- Time to first answer token
- 282.51sConfiguration: Prompt length: 1,000 · Parallel queries: 1Artificial Analysis · Date unknownAge unknown
API pricing
- API input cost
- API output cost
Benchmark variant: Claude Sonnet 5.5 (Max, Default Fallback)
AA data retrieved Nov 2, 2026 · Artificial Analysis
Indexes use points; evaluations use accuracy percentages. API measurements do not measure yno application performance. Methodology
Capabilities
- Reasoning
- Code
- Agentic
- Tool Use
- Long Context
- Vision
Use cases
- Agentic software engineering
- Browser & terminal automation
- Full-stack development
- Production AI agents
Strengths
- Vendor Terminal-Bench 4.0 70.6% (up from 10.3%)
- ~30% faster and ~30% cheaper per task than Sonnet 5
- 1M context / 128K output at $2/$10 per MTok
- First Sonnet under Opus-class cyber safeguards
Best for
Teams that want a fast, cost-efficient Claude 5.5 model for coding and agent workflows
No credit card required
Related models
Frequently asked
- What is Claude Sonnet 5.5?
- Anthropic's September 28, 2026 model and the second in the Claude 5.5 family after Opus 5.5, built for coding and agent workflows at Sonnet pricing: $2 input and $10 output per million tokens, unchanged from Sonnet 5, with cache reads at $0.10 (cut from $0.20 in October 2026) and cache writes at $2.50. It accepts text and image input, provides a 1M-token context window and up to 128K output, and keeps thinking on with adjustable effort (medium default in Claude apps, high on the Platform). Anthropic reports roughly 30% faster and about 30% cheaper per task than Sonnet 5, Terminal-Bench 4.0 of 70.6% (up from Sonnet 5's 10.3%) and near-Opus-5.5 results on GDPval-AA, and it is the first Sonnet released under Opus-class cyber safeguards. The benchmark figures above are Anthropic's own; Artificial Analysis measurements are shown separately.
- How much does Claude Sonnet 5.5 cost in yno.ai?
- Claude Sonnet 5.5 is available on the Pro tier of yno.ai.
- What can Claude Sonnet 5.5 do?
- Claude Sonnet 5.5 is best for Teams that want a fast, cost-efficient Claude 5.5 model for coding and agent workflows. Its main capabilities include Reasoning, Code, Agentic, Tool Use, Long Context, Vision.
- How does Claude Sonnet 5.5 compare to other models?
- Claude Sonnet 5.5 excels at Vendor Terminal-Bench 4.0 70.6% (up from 10.3%) and is recommended for Agentic software engineering, Browser & terminal automation, Full-stack development, Production AI agents. See related models below.
- How do I use Claude Sonnet 5.5 in yno.ai?
- Sign up for yno.ai, select Claude Sonnet 5.5 from the model picker, and start chatting.