Anthropic: Claude Sonnet 4.6

Claude Sonnet 4.6 is a closed, paid model from Anthropic that accepts text, image, and file inputs and supports both tool use and reasoning. Its context window reaches one million tokens, and completions can run up to 128,000 tokens, making it well-suited for long-document work, multi-step agentic pipelines, and tasks that combine code, vision, and file analysis in a single session. At $3.00 per million input tokens and $15.00 per million output tokens, it sits in the mid-to-upper tier of hosted model pricing. Its blended benchmark score of 84.4 covers five benchmarks, including a perfect 100 on the agentic index and 84.1 on coding, so teams running tool-heavy or agentic workflows have the most direct evidence in its favor. General reasoning and emotional-intelligence scores are solid but not exceptional. Buyers who prioritize large context and agentic reliability will find the data reasonably persuasive; those optimizing for cost or evaluating pure conversational tasks should compare it against lower-priced alternatives with similar coverage.