Claude Sonnet 5.5 Launches With Lower Task Costs and Cursor Support
Anthropic introduced Claude Sonnet 5.5 on September 28 with model ID `claude-sonnet-5-5`. It keeps Sonnet 5's $2 input and $10 output price per million tokens; Anthropic reports output generation more than 30% faster and up to 30% lower cost per task. Cursor lists the model in Settings > Models. Cache reads dropped to $0.10 per million tokens on October 7.
Anthropic introduced Claude Sonnet 5.5 on September 28 as the second model in its Claude 5.5 family. The model keeps Sonnet 5's API prices at $2 per million input tokens and $10 per million output tokens; Anthropic says its output generation is more than 30% faster and that it costs up to 30% less per task in its testing (Anthropic).
Sonnet 5.5 is positioned for well-scoped coding and knowledge-work tasks where the result can be checked. Anthropic describes Opus 5.5 as the better fit for complex, open-ended work requiring sustained judgment (Anthropic's model announcement; Claude's model-selection guide).
Key Takeaways
- Model ID:
claude-sonnet-5-5on Anthropic's developer platform (Anthropic). - Token price: $2 per million input tokens and $10 per million output tokens, the same as Sonnet 5 (Anthropic).
- Cache reads: $0.10 per million tokens since October 7, 2026, 0.05x the base input price, down from $0.20 (Anthropic release notes; Anthropic pricing).
- Task efficiency: Anthropic reports output generation more than 30% faster and up to 30% lower cost per task than Sonnet 5 in its testing (Anthropic).
- Benchmarks: Anthropic reports 70.6% on Terminal-Bench 4.0, compared to Sonnet 5's 10.3% (Anthropic).
- Cursor: Cursor lists Sonnet 5.5 in its model picker. Add it from Cursor Settings > Models (Cursor model documentation).
- Claude Code settings: Sonnet 5.5 requires Claude Code v2.1.284 or later (Claude Code model configuration). Claude Code and Claude apps default to Medium effort; the Claude Platform defaults to High (Anthropic).
What Sonnet 5.5 Is For
Anthropic recommends Sonnet 5.5 for well-scoped everyday coding, bug fixes, feature iteration, and polished documents, slides, and spreadsheets. Its developer guide says the model fits best when a task has a clear specification and a way to check the result. For the hardest long-horizon work, Anthropic recommends considering Opus 5.5 instead (Anthropic; Claude's model-selection guide).
The speed and cost claims refer to work completed, not a lower per-token price. Sonnet 5.5 keeps the same listed input and output rates as Sonnet 5, while Anthropic says it typically uses fewer tokens to do the same work. Its announcement reports output generation more than 30% faster and up to 30% less cost per task in Anthropic's testing (Anthropic).
On October 7, 2026, Anthropic lowered the price of prompt cache reads on Sonnet 5.5 from $0.20 to $0.10 per million tokens, 0.05x the base input price instead of 0.1x. Cache writes and all other prices are unchanged (Anthropic release notes; Anthropic pricing).
On benchmarks, Anthropic says Sonnet 5.5 scores 70.6% on Terminal-Bench 4.0, an agentic coding evaluation, compared to Sonnet 5's 10.3%, and scores two points below Opus 5.5 on GDPval-AA, a test of real-world work across a variety of occupations (Anthropic).
Where to Use It
Anthropic lists claude-sonnet-5-5 as the model identifier for developers starting on the Claude Platform. Its announcement also says Sonnet 5.5 is available on Amazon Web Services, Google Cloud, and Microsoft Azure (Anthropic; Anthropic's migration guide).
Cursor's model documentation lists Claude Sonnet 5.5 as an available model and directs users to Cursor Settings > Models to add it to the model picker. The page says Sonnet 5.5 supports thinking mode and context windows up to 1M tokens (Cursor model documentation). Sonnet 5.5 draws from Cursor's third-party Other Models usage pool. Cursor lists global-endpoint prices of $2 per million input tokens and $10 per million output tokens (Cursor model documentation).
In Claude Code, Sonnet 5.5 requires v2.1.284 or later; Anthropic's docs say to run claude update to upgrade (Claude Code model configuration).
Effort Settings and API Migration
Effort defaults differ by surface: Anthropic lists Medium for Claude Code and the Claude apps, and High for the Claude Platform. Anthropic's developer guide recommends starting at High for general API use, Medium for well-specified agentic coding tasks, and Medium or Low for latency-sensitive chat. Effort levels are recalibrated, so the same label does not necessarily produce the same amount of thinking as it did on Sonnet 5 (Anthropic; Claude's model-selection guide).
If an API integration used thinking: {"type": "disabled"} on Sonnet 5, Anthropic says that setting is not supported on Sonnet 5.5. The migration guide says to use thinking: {"type": "between_tools"} to turn off up-front thinking; this setting is accepted at Low, Medium, and High effort, but not at Xhigh or Max. Anthropic also recommends rerunning effort tests and re-baselining cost after a migration (Anthropic's migration guide).
Anthropic's release notes list five ways code written for Sonnet 5 can break on Sonnet 5.5: the "disabled" thinking type gives way to "between_tools", forced tool use (tool_choice types any and tool) returns a 400 error, thinking blocks are tied to the model and the conversation, the Claude API and Google Cloud no longer accept the earlier computer_20251124 tool, and the advisor tool rejects Claude Opus 4.8, Claude Opus 4.7, and Claude Sonnet 5 as advisors. Our Sonnet 5.5 prompting guide walks through each change (Anthropic release notes).
Sources
- Anthropic, “Introducing Claude Sonnet 5.5,” September 28, 2026: https://www.anthropic.com/claude-sonnet-5-5
- Anthropic, “Migrating to Claude Sonnet 5.5”: https://platform.claude.com/docs/en/models/sonnet-5-5/migration-guide
- Claude Devs, “Building with Claude Sonnet 5.5,” September 28, 2026: https://claude.dev/blog/building-with-claude-sonnet-5-5/
- Cursor Docs, “Claude Sonnet 5.5”: https://cursor.com/docs/models/claude-sonnet-5-5
- Anthropic, Claude Platform release notes, October 7, 2026 (cache-read price cut) and September 28, 2026 (launch and breaking changes): https://platform.claude.com/docs/en/release-notes/overview
- Anthropic, Claude Platform pricing, checked October 9, 2026: https://platform.claude.com/docs/en/about-claude/pricing
- Anthropic, Claude Code docs, "Model configuration", checked October 9, 2026: https://code.claude.com/docs/en/model-config
Read next
Related coverage- Claude
Prompting Claude Sonnet 5.5: Effort, between_tools, and the Five Breaking Changes
Run Claude Sonnet 5.5 at high effort for general API work, medium for well-specified agentic coding, and medium or low for chat. Thinking is on by default; turn it off with between_tools, not disabled. Five API changes break Sonnet 5 code, and cache reads now cost $0.10 per million tokens.
- Claude
Claude Opus 5.5: Fable 5.1-Level Scores at $4 and $20, With Benchmarks, Effort Costs, and Migration Guide
Claude Opus 5.5, released September 22, 2026, beats Claude Fable 5.1 on every benchmark Anthropic published and costs $4 and $20 per million tokens, 60% less than Fable 5.1. It scores 66.4% on Terminal-Bench 4.0 against 57.9% for GPT-6 Astra. Thinking is always on, the default effort is medium, and four API changes can break Opus 5 code.
- Claude Code
Claude Code Projects Turn One Conversation Into Parallel Cloud Sessions
Anthropic's redesigned Projects beta gives Claude Code one coordinating conversation that delegates work to parallel threads. Threads are usually cloud sessions, which Anthropic's docs now list as available, no longer in research preview, on Pro, Max, and Team and for eligible Enterprise seats. A project can also run a thread on your own machine through Remote Control.
Frequently Asked Questions
How much does Claude Sonnet 5.5 cost?
Claude Sonnet 5.5 costs $2 per million input tokens and $10 per million output tokens, the same as Sonnet 5. Since October 7, 2026, prompt cache reads cost $0.10 per million tokens, 0.05x the base input price, down from $0.20. Cache writes and all other prices are unchanged.
Which Claude Code version supports Sonnet 5.5?
Sonnet 5.5 requires Claude Code v2.1.284 or later, according to Anthropic's model configuration docs, and Opus 5.5 requires v2.1.280 or later. Run claude update to upgrade if a request for Sonnet 5.5 fails. Claude Code and the Claude apps run Sonnet 5.5 at Medium effort by default.
Will code written for Sonnet 5 break on Sonnet 5.5?
It can, in five ways. The disabled thinking type is replaced by between_tools at high effort or below, forced tool use returns a 400 error, thinking blocks are tied to the model and the conversation, computer_20251124 is not accepted on the Claude API and Google Cloud, and the advisor tool rejects older advisors.
How does Sonnet 5.5 score on benchmarks?
Anthropic reports that Sonnet 5.5 scores 70.6% on Terminal-Bench 4.0, an agentic coding evaluation, compared to Sonnet 5's 10.3%. It scores two points below Opus 5.5 on GDPval-AA, a test of real-world work across occupations. Anthropic also reports output generation more than 30% faster than Sonnet 5.