Google: Gemini 4 Carbon

Gemini 4 Carbon is a Google model that accepts audio, file, image, text, and video inputs, so one deployment can cover mixed-media workloads. It offers a context window of 1,048,576 tokens and allows up to 943,718 completion tokens, which suits long documents, large codebases, or lengthy transcripts in a single request. It supports tool calling and reasoning; structured output support is not specified in the available data. At $2.00 per million input tokens and $10.00 per million output tokens, it is a paid model, and output-heavy jobs will cost noticeably more than input-heavy ones. Teams that need multimodal input, very long context, and tool use should shortlist it. Its blended benchmark score of 98.9 is strong but comes from a single independent benchmark (AI Index, 99.0), so treat its standing as provisional and test it on your own tasks before committing.