模型多源确认

Anthropic Sonnet 5.5模型表现不佳

精选理由

Anthropic新发布的Sonnet 5.5模型表现不如预期,还不如旧版Sonnet 4.6和DeepSeek Flash。

Anthropic最新发布的Sonnet 5.5模型在最大模式下表现不佳,在智能体编程任务中性能欠佳,表现略逊于Terra模型。用户建议继续使用Sonnet 4.6或DeepSeek Flash模型。Anthropic应考虑淘汰Sonnet和Haiku模型,发布其大型Model-2模型。

原文 · Bindu Reddy

Sonnet 5.5 Is Not A Very Good Model

- spins on max mode - underperforms on agentic coding - slightly worse than Terra - stick either Sonnet 4.6 or DeepSeek Flash

Anthropic should retire Sonnet and Haiku, release their big Model-2