← All models

Claude 3.7 Sonnet

Published Publisher Anthropicclaude-3-7-sonnet-20250219

Model Positioning and Release Date

Claude 3.7 Sonnet is a hybrid reasoning model released by Anthropic on February 24, 2025, offering both standard responses and extended thinking in the same model. (anthropic.com)

Capabilities and Uses

  • Supports text and image input and produces text output; it can be used for conversations, content analysis, creative tasks, and programming. (anthropic.com)
  • Extended thinking mode performs deeper reasoning before responding; API users can set a thinking token budget to balance speed, cost, and response quality. Anthropic says it is suitable for tasks such as mathematics, physics, instruction following, and coding. (anthropic.com)
  • Can be used for code generation, codebase work, and other business tasks requiring multi-step analysis. (anthropic.com)

Limitations and Service Status

  • Knowledge cutoff: October 2024; the model's own knowledge of subsequent events may be incomplete. (anthropic.com)
  • Outputs may be incorrect, and important results should be checked by a human; extended thinking does not guarantee correct answers either.
  • The model was retired from the Claude API on February 19, 2026; ordinary requests will return an error. Anthropic states that researchers can apply for continued access through the external researcher access program. (platform.claude.com)

Pricing

At launch, Anthropic announced the same pricing for standard and extended thinking modes: 3 US dollars (USD) per 1 million input tokens and 15 US dollars (USD) per 1 million output tokens; output billing includes thinking tokens. These are historical prices from the launch announcement and do not indicate that the model can be purchased or accessed normally through the Claude API after retirement. Official pricing source: launch announcement; date accessed: November 19, 2026. (anthropic.com)

Official Sources