Anthropic releases Sonnet 5.5, which it calls a significantly cheaper, faster work partner

Anthropic unveiled Sonnet 5.5, the newest version of its mid-range assistant model, positioning it as a faster and more cost-efficient option for everyday tasks such as coding and office document creation. The company highlights speed gains as the primary improvement: Sonnet 5.5 is said to run roughly 30% faster than Sonnet 5 and to consume tokens at a noticeably lower rate.

Sonnet occupies a middle ground in Anthropic’s model lineup, below the more powerful Opus models but designed for agility and lower-cost agentic deployments. According to Anthropic’s internal benchmarks, Sonnet 5.5 outperforms Opus 5.5 on agentic coding tasks, a result the company attributes to Sonnet’s ability to spawn multiple agents without driving costs above practical limits.

Anthropic also claims that Sonnet 5.5 has cyber capabilities comparable to Opus 5. Because of that parity, Sonnet 5.5 will be covered by the same cyber safeguards that the company already applies to its Fable and Opus models. The firm said it plans to release an updated Haiku model — its smallest offering — in the coming weeks, but did not provide a specific launch date.

The release of Sonnet 5.5 arrives amid a surge of model launches across the industry. In the past week, OpenAI introduced updated versions of its mid-tier and budget models Sol and Luna, and Meta announced a new model intended to support a forthcoming smart-glasses feature. This report is based on coverage by Lucas Ropek at TechCrunch.