Anthropic: Claude Sonnet 4.6

Claude Sonnet 4.6 is Anthropic's mid-priced model, accepting text, image, and file inputs with a 1,000,000-token context window and up to 128,000 output tokens per response. It supports tool calling and reasoning, which suits it to agentic workflows and multi-step tasks; structured output support is not specified in the available data, so teams that depend on strict schema guarantees should verify that before committing. At $3 per million input tokens and $15 per million output tokens, it is not a budget option, and output-heavy workloads will cost noticeably more than input-heavy ones. Its blended benchmark score of 80.1 across five independent benchmarks gives a reasonably grounded picture. Coding is the clear strength, with a top coding index score of 100 and 61.3 on Aider Polyglot, while its agentic index of 67.3 and EQ-Bench score of 68.4 are more moderate. Developers weighing code-focused or long-document work should shortlist Claude Sonnet 4.6 and compare it against cheaper alternatives on their own tasks.