Today, AI company Anthropic officially launched its latest "hybrid reasoning model"—Claude 3.7 Sonnet. Anthropic hails this model as its "most intelligent" AI yet, designed to tackle more complex problems and exhibiting superior performance in areas like mathematics and coding.
According to Anthropic, Claude 3.7 Sonnet, as a "hybrid" model, combines a "reasoning mode" (able to pause and consider complex answers) with a traditional mode (generating answers in real-time). This is currently the only such "hybrid" model on the market. It enhances the ability to handle complex problems while maintaining fast response times, resulting in little difference in speed whether answering simple questions like "What time is it?" or complex prompts like "Plan a two-week trip to Italy, considering the weather in late March."
Beyond the new model, Anthropic also released a "limited research preview" of its autonomous programming tool, Claude Code. Positioned as a "proactive partner," Claude Code can search and read code, edit files, write and run tests, commit code to GitHub, and use command-line tools. This expands the application scenarios for the Claude family of models.
Starting this week, users can access Claude 3.7 Sonnet within the Claude app, and developers can access it through Anthropic's API, Amazon Bedrock, and Google Cloud's Vertex AI. The cost is the same as its predecessor, 3.5 Sonnet: $3 per million input tokens and $15 per million output tokens.
Furthermore, the model's knowledge cutoff date is October 2024, making it more up-to-date than other models.
Notably, Anthropic allows developers to guide the model's "thinking" process using its "scratchpad" feature, even precisely controlling response times. This provides developers with greater flexibility and control, allowing Claude 3.7 Sonnet to better adapt to various applications.
