Release2026-09-28
Anthropic released Claude Sonnet 5.5 as the second model in the Claude 5.5 family. It runs 30%+ faster than Sonnet 5. It costs up to 30% less per task than Sonnet 5. It scores 70.6% on Terminal-Bench 4.0 versus 10.3% for Sonnet 5. It matches Opus 5.5 on some knowledge work benchmarks. It is the first Sonnet model with cyber safeguards like Opus 5. It keeps the same pricing as Sonnet 5. It is suited for well-scoped everyday tasks and design work. Claude Haiku 5.5 will join the family in the coming weeks. It is available now.
What changed
Claude Sonnet 5 was the prior model in this tier.
What it unlocks
Running everyday coding and knowledge tasks faster and cheaper than before.
- 70.6% on Terminal-Bench 4.0
- 30%+ faster output
- up to 30% cheaper per task
- $2 per 1M input tokens
Sources