anthropic/claude-sonnet-4.6

Claude Sonnet 4.6 from Anthropic: a full upgrade to coding, computer use, long-context reasoning, agent planning, knowledge work, and design, with a 1 million token context window in beta.

4 runs
Playground API Examples README

Claude Sonnet 4.6

Claude Sonnet 4.6 is Anthropic’s latest Sonnet model, released February 17, 2026. It’s a full upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design — and it ships with a 1 million token context window in beta.

In early testing, developers preferred Sonnet 4.6 to Sonnet 4.5 by a wide margin, and even preferred it to Anthropic’s larger Opus 4.5 model 59% of the time. Customers describe it as less prone to overengineering, better at following instructions, more honest about what it has and hasn’t done, and easier to use across long sessions.

What it’s good at

  • Coding. In Claude Code, early users preferred Sonnet 4.6 to Sonnet 4.5 roughly 70% of the time. It reads context before modifying code, consolidates shared logic instead of duplicating it, and follows multi-step instructions more consistently.
  • Computer use. Sonnet 4.6 is a major step up over Sonnet 4.5 on OSWorld-Verified, the standard benchmark for navigating real software like Chrome, LibreOffice, and VS Code by clicking and typing the way a person would. It’s also significantly more resistant to prompt injection attacks during computer use.
  • Long-context reasoning. The 1 million token context window is enough to hold an entire codebase, lengthy contracts, or dozens of research papers in a single request. Sonnet 4.6 reasons effectively across that context, which makes it much better at long-horizon planning.
  • Frontend and design. Customers report that Sonnet 4.6 produces more polished visual output — better layouts, animations, and design sensibility — and needs fewer rounds of iteration to reach production-quality results.
  • Document comprehension. Sonnet 4.6 matches Opus 4.6 on Databricks’ OfficeQA, which measures how well a model reads enterprise documents (charts, PDFs, tables) and reasons over them.

Inputs

  • prompt — the user message.
  • system_prompt — optional system prompt.
  • image — optional input image. Images are priced as (width px * height px) / 750 input tokens.
  • max_tokens — maximum number of output tokens, up to 64,000.
  • max_image_resolution — caps input image resolution in megapixels to save time and money. Defaults to 0.5 MP.

Pricing

Sonnet 4.6 is priced the same as Sonnet 4.5: $3 per million input tokens, $15 per million output tokens. See the Replicate pricing on the right of this page for details.

Model created