Anthropic Launches Claude Sonnet 5.5 with Enhanced Cyber Safeguards

• •Sonnet 5.5 scores 70.6% on Terminal-Bench 4.0, outperforming Sonnet 5's 10.3%.
• •The model features enhanced cyber safeguards and classifiers to block reasoning extraction.
• •Sonnet 5.5 is priced at $2 per million input tokens and $10 per million output tokens.

Anthropic has released Claude Sonnet 5.5, which boasts a 70.6% score on the Terminal-Bench 4.0 agentic coding test, significantly outperforming its predecessor Sonnet 5, which scored 10.3%. The new model is designed to be 30% faster and 30% cheaper per task than Sonnet 5, priced at $2 per million input tokens and $10 per million output tokens. Notably, Sonnet 5.5 includes frontier-style cyber safeguards and classifiers that block reasoning extraction, enhancing its cybersecurity capabilities to be comparable to the more expensive Opus 5.5 model. The model's launch follows new European regulations requiring enhanced cybersecurity measures for AI models with systemic risks. Anthropic claims that while Sonnet 5.5 does not advance the frontier of its models' capabilities, it still warrants significant cyber safeguards. The company also plans to release a new version of its smallest model, Haiku, in the near future.