Anthropic releases Claude 3.7 Sonnet and previews Claude Code

Anthropic/Claude models

Anthropic released Claude 3.7 Sonnet, a hybrid reasoning model that can either answer instantly or think step-by-step for longer, with API users able to set computation budgets. Pricing stayed at $3/$15 per million input/output tokens. Alongside it came Claude Code, a command-line tool for delegating programming tasks, released as a limited research preview.

• Claude 3.7 Sonnet: first hybrid reasoning model; users choose speed vs. depth; thinking up to 128K tokens
• Pricing unchanged: $3 per million input tokens, $15 per output tokens (including thinking tokens)
• State-of-the-art on SWE-bench Verified and TAU-bench coding benchmarks
• Claude Code: research preview; command-line tool for delegating programming tasks; searches code, edits, tests, pushes to GitHub