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 maximum output of 131,072 tokens. It supports tool use and reasoning, making it capable of agentic workflows and multi-step tasks. Structured output support is unconfirmed. The model is not free and not listed as open weights. 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 64.8 is drawn from only three benchmarks, so treat that figure as preliminary. The coding sub-score of 86.8 is notably high and worth attention for developer use cases, while the agentic score of 42.2 is comparatively low, suggesting caution for complex autonomous task pipelines. Budget-conscious teams running code-heavy workloads have the clearest reason to shortlist it.