Anthropic has released Sonnet 5.5, a new version of its mid-tier AI model that the company said is faster and less expensive than Sonnet 5.

Anthropic said the model is aimed at everyday tasks such as coding and creating office documents.

The company said Sonnet 5.5 is 30% faster than the earlier model and has a slower token burn rate. In Anthropic’s benchmarks, Sonnet 5.5 also performed better than Opus 5.5 on agentic coding.

Anthropic also said Sonnet 5.5 has cyber capabilities comparable to Opus 5. That makes it the first Sonnet model subject to the same safeguards Anthropic applies to Fable and Opus.

Anthropic has outlined its approach to handling more capable AI systems in its Responsible Scaling Policy, though the available policy materials did not specify how an “Opus-level” cyber designation maps to particular deployment restrictions.

The company said it plans to release an updated Haiku model in the coming weeks, as OpenAI and Meta continue rolling out new models.

🔗 Source: Anthropic

Recent Anthropic developments