Model Overview & Capabilities
Claude Sonnet 5.5 is Anthropic’s balanced frontier coding and agentic execution model officially launched on September 28, 2026, holding the premier Rank #2 position on our leaderboard with an empirical composite score of 55.3. In standardized independent evaluations, Sonnet 5.5 achieved a chart-topping 70.6% on Terminal-Bench 2.0 alongside 50.5% on SWE-bench Verified, 58.0% on Aider Polyglot, and 52.5% on LiveCodeBench, decisively outperforming the previous generation across all practical software development tasks.Available at an attractive API price of $2.00 per million input tokens and $10.00 per million output tokens ($4.00 blended price at 3:1 ratio), Sonnet 5.5 provides remarkable daily engineering value, achieving an industry-leading $0.0723 per index point. The model pairs this high efficiency with rapid generation throughput at 136 tokens per second and an exceptionally responsive 1.16-second time to first token.
Equipped with a unified 1,000,000-token context window, 64,000 maximum generation tokens, and prompt caching at $0.10 per million tokens, Sonnet 5.5 is exceptionally well-suited for interactive pair programming, continuous integration code reviews, automated test suite maintenance, and high-concurrency developer tooling inside leading development environments.
