拡張思考に対応した「Claude 3.7 Sonnet」

Anthropicが「Claude 3.7 Sonnet」を発表しました。Anthropicによれば、Claude 3.7 Sonnetは「市場初のハイブリッド推論モデル」で、ベンチマークではOpenAIのo1やo3-mini、DeepSeek-R1を超える性能とのことです。
OpenAIとGoogleの2強対決の様相を呈していたのがここ２～3カ月でしたが、やっとAnthropicが推論モデルをリリースし、Grok 3と合わせて、4強の戦いとなるのか今後も注目です。
 
 
Claude 3.7 Sonnet and Claude Code
2025年2月25日
https://www.anthropic.com/news/claude-3-7-sonnet
 
Claude’s extended thinking
2025年2月24日
https://www.anthropic.com/research/visible-extended-thinking
 
 
"Claude 3.7 Sonnet" Adapted for Extended Thinking
 
Anthropic has announced the release of Claude 3.7 Sonnet. According to Anthropic, Claude 3.7 Sonnet is "the market's first hybrid reasoning model" and has outperformed OpenAI’s o1 and o3-mini, as well as DeepSeek-R1, in benchmark tests.
Over the past two to three months, the AI landscape had largely been a two-horse race between OpenAI and Google. However, with Anthropic finally launching its new reasoning model and Grok 3 also in the mix, it remains to be seen whether this will now evolve into a four-way competition.