MiniMax
Models published by MiniMax, available through the AnyRouter API. Each can route across multiple upstream providers for availability and price.
MiniMax M2.5 is a SOTA large language model designed for real-world productivity. Trained in diverse digital working environments, M2.5 extends M2.1 coding expertise into general office work — generating and operating Word, Excel, and PowerPoint files, context switching across software, and collaborating across agent and human teams. Scores 80.2% SWE-Bench Verified, 51.3% Multi-SWE-Bench, and 76.3% BrowseComp.
MiniMax M2.7 is a large language model for complex software engineering, agentic tool use, and office productivity workflows. Supports complex agent harnesses, dynamic tool search, Agent Teams, and high-fidelity coding and document-editing tasks. Available via NVIDIA NIM and Cloudflare AI Gateway.
MiniMax M2 is a large language model for complex software engineering, agentic tool use, and office productivity workflows. Supports complex agent harnesses, dynamic tool search, Agent Teams, and high-fidelity coding and document-editing tasks.
MiniMax's M3 language model with frontier coding and agentic capabilities, a 1M token context window, and multilingual support.
MiniMax-M3.1-Flash is a multimodal model from MiniMax. Compared with M3, it improves across benchmarks, particularly complex agentic tasks, coding and cybersecurity. It is built for multi-step co-work automation, understanding real codebases, fixing bugs and building features, and for visual understanding and creation, including frontend and presentation work. It has a 1M token context window and always-on reasoning with adjustable effort.
MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity with advanced agentic capabilities through multi-agent collaboration. Features strong reasoning, function calling, and structured output capabilities with 198K context window.