Xiaomi: MiMo-V2.5-Pro

MiMo-V2.5-Pro is a text-only model from Xiaomi with a 1-million-token context window and a 131,072-token output ceiling. It supports tool use and reasoning, making it suitable for multi-step workflows and agentic tasks, though structured output support is unconfirmed. No image or audio input is available, so document or media pipelines will need a different solution. At $0.435 per million input tokens and $0.87 per million output tokens, it sits in the budget-to-mid tier for reasoning-capable models. Its blended benchmark score of 74.0 is drawn from only three benchmarks, so treat that figure as a limited signal rather than a settled verdict. The coding sub-score of 99.3 is notable and worth testing for code-heavy use cases, while the agentic score of 48.0 suggests caution for fully autonomous agent workflows. Teams running high-volume coding tasks on a cost-sensitive budget have the clearest reason to shortlist it.