Anthropic Updates Sonnet 5.5: Prices Unchanged, But More Token-Efficient for the Same Tasks

Beating AI Newsflash: Anthropic has released Claude Sonnet 5.5, continuing to focus on everyday tasks and programming. The API pricing remains the same as Sonnet 5, with input and output costs at $2 and $10 per million tokens, respectively—exactly half the price of Opus 5.5. Anthropic claims the new model outputs over 30% faster and typically requires fewer tokens to complete the same tasks, potentially reducing per-task costs by up to 30%. The most significant improvement is seen in agent programming. In Terminal-Bench 4.0 (which tests models on multi-step tasks in the command line), Sonnet 5.5 jumped from Sonnet 5's 10.3% to 70.6%, surpassing Opus 5.5's 66.4%. However, among the eight benchmarks officially released, Sonnet 5.5 only outperformed Opus in this one category, while Opus 5.5 still led in the other seven, though the gaps in most cases have narrowed significantly. Anthropic also explicitly stated that complex, open-ended tasks requiring sustained judgment are still better suited for Opus 5.5. But when calculated on a per-task basis, the cost savings may not necessarily reach 50%. In the highest Max reasoning mode tested by Artificial Analysis, Sonnet 5.5 achieved an Intelligence Index of 56, compared to Opus 5.5 Max's 58. However, Sonnet 5.5 used an average of approximately 193,000 output tokens per task in this test, resulting in a per-task cost of $7.60—higher than Opus 5.5 Max's $5.98. Whether it actually saves money depends heavily on the task and reasoning mode. Sonnet 5.5 is now available on Claude, Claude API, AWS, Google Cloud, and Microsoft Foundry. Haiku 5.5 will be released in the coming weeks. [Original Link]