Anthropic has rolled out Claude Sonnet 5.5 as the second release in its Claude 5.5 model family. According to MarkTechPost, the new model achieves a score of 70.6% on Terminal-Bench 4.0 and comes within two points of Opus 5.5 on GDPval-AA. Anthropic states that Claude Sonnet 5.5 runs over 30% faster than its predecessor, Sonnet 5, while maintaining pricing of $2 per million input tokens and $10 per million output tokens. The company also claims that overall cost per task drops by up to 30% because the model uses fewer tokens to complete requests. Availability spans the Claude API alongside cloud platforms including Amazon Web Services, Google Cloud, and Microsoft Azure.
Performance And Cost Efficiency
The release marks a continuation of Anthropic’s iterative updates to its mid-tier model lineup. By matching the previous pricing structure while improving execution speed and reducing token consumption, the company is targeting higher efficiency for automated workflows and developer tasks. Benchmarking data shows strong performance on Terminal-Bench 4.0, placing Sonnet 5.5 close to the larger Opus 5.5 variant on specific evaluations.
What Remains Unclear
Despite the release details shared by Anthropic, several operational specifics are absent from the announcement. The documentation and public disclosures do not specify the exact parameter count, hardware requirements, or context window limits for Claude Sonnet 5.5. Additionally, the announcement did not detail a specific timeline for broader consumer availability outside of the listed enterprise cloud platforms and API options.
Developers can begin deploying Claude Sonnet 5.5 immediately across supported cloud environments and the official API.
Lead image: Brett Sayles / Pexels
