MiniMax: MiniMax M2.7

MiniMax M2.7 is a text-in, text-out model from MiniMax with a 204,800-token context window and a 131,072-token output limit. It supports tool use and reasoning, making it suitable for multi-step agentic workflows and tasks that benefit from structured intermediate thinking. Structured output support is unconfirmed, so teams with strict JSON requirements should verify before committing. At $0.25 per million input tokens and $1.00 per million output tokens, MiniMax M2.7 sits at the affordable end of the market. Its blended benchmark score of 75.9 covers only three benchmarks, so confidence in that figure is limited compared to more widely tested models. The standout data point is a perfect score on the agentic benchmark, which may attract developers building tool-calling pipelines on a budget. General reasoning and coding scores are moderate, so users with demanding accuracy requirements should weigh those numbers carefully.