Anthropic launched Claude Sonnet 5.5, the second model in the Claude 5.5 family. Per the official announcement, it is a clear upgrade over Sonnet 5: more than 30% faster, and up to 30% cheaper per typical task, without changing list prices. Pricing stays $2 per million input tokens, $10 per million output tokens, and $0.20 for cache reads, but the model usually needs fewer tokens for the same work.
For current plans, prices and every Maitz article on Claude, see our Claude hub.
Sonnet 5.5 is meant to complement Opus 5.5: Opus for complex work that needs careful judgment, Sonnet for well-scoped everyday tasks, bug fixes, documents, slides and spreadsheets, and design polish. Anthropic says Haiku 5.5, aimed at high-volume and cost-sensitive use, will join the family in the coming weeks. More updates in AI News.
What Anthropic reports on performance
These figures are Anthropic’s own reported numbers from the announcement, not independent verification. On Terminal-Bench 4.0 (an agentic coding evaluation), Sonnet 5.5 scores 70.6%, versus Sonnet 5’s 10.3%. On GDPval-AA, a test of real-world work across occupations, it is nearly level with Opus 5.5: 1844 versus 1846. The company stresses that benchmarks are only one slice of capability, and that Opus 5.5 remains stronger on open-ended work that needs sustained judgment.
Speed, cost, and availability
Beyond lower token use, Anthropic says Sonnet 5.5 generates output more than 30% faster than Sonnet 5. It is available on all platforms, including AWS, Google Cloud, and Azure. The API id is claude-sonnet-5-5. Like Opus 5.5 and Sonnet 5, it ships with zero data retention. If you run Sonnet with thinking off, switch to the between_tools setting before moving to 5.5.
Safety and safeguards
According to Anthropic, Sonnet 5.5’s cybersecurity capability is comparable to Opus 5’s, so this is the first Sonnet to launch with cyber safeguards and fallbacks like those on the stronger models. Biology safeguards match Sonnet 5. It also adds distillation classifiers and preserved thinking, so Claude’s thinking cannot be decoupled from the account that created it. Routine software development and most life-sciences work are meant to stay unaffected by the narrow high-risk restrictions.
For Claude in orgs and developer tooling, see also Barclays expanding Claude, Claude Frontier Academy, and Claude Code TypeScript mods.
Sources
- Anthropic, Introducing Claude Sonnet 5.5, official announcement: anthropic.com/claude-sonnet-5-5 (September 28, 2026)
