Anthropic's New Coding Model Beats Opus 5.5 on Two Benchmarks, but Still Recommends Opus

Anthropic released a coding model that outperforms Opus 5.5 on two major benchmarks and can complete certain difficult tasks. Despite its stronger results, Anthropic recommends customers continue using Opus. The source excerpt gives few details on the model's name, pricing, or the reasoning behind the recommendation.

Why it matters: Developers choosing between Anthropic models should weigh benchmark gains against the company's own guidance, which may reflect cost or reliability trade-offs.