MiniMax M2

2025-10-27 · Source: Ollama Blog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, quick

Summary

MiniMax M2, a new model optimized for coding and agentic workflows, is now available on Ollama's cloud as of October 28, 2025. This model features 10 billion activated parameters (230 billion total) and is designed for efficiency, offering lower latency, reduced cost, and higher throughput for interactive agents and batched sampling. Benchmarks from Artificial Analysis indicate MiniMax-M2 achieves a #1 composite score among open-source models globally for general intelligence across mathematics, science, instruction following, coding, and agentic tool use. It excels in multi-file edits, coding-run-fix loops, and test-validated repairs, demonstrating strong performance on Terminal-Bench and (Multi-)SWE-Bench-style tasks. Furthermore, MiniMax-M2 shows robust agent performance, planning and executing complex, long-horizon toolchains across various environments like shell, browser, retrieval, and code runners.

Key takeaway

For engineering leaders and developers evaluating AI models for coding and agentic tasks, MiniMax M2 presents a compelling option. Its #1 global ranking among open-source models for general intelligence and specialized performance in multi-file edits and complex agentic workflows suggests it can significantly enhance developer productivity. Consider integrating MiniMax M2 into your team's IDEs and CI pipelines to capitalize on its efficiency and advanced coding capabilities, potentially reducing operational costs and improving throughput for interactive AI agents.

Key insights

MiniMax M2 is a top-ranked open-source model optimized for coding and agentic workflows with high efficiency.

Principles

• Efficiency enhances interactive agent and batched sampling performance.
• Agentic models require robust planning and error recovery capabilities.

Method

MiniMax M2 can be integrated into developer environments like VS Code, Zed, and Droid by pulling the model via `ollama pull minimax-m2:cloud` and configuring the respective IDE's LLM provider settings.

In practice

• Use `ollama run minimax-m2:cloud` to get started.
• Integrate with VS Code, Zed, or Droid for coding tasks.
• Access via Ollama's Cloud API using an API key.

Topics

• MiniMax M2
• Code Generation
• Agentic AI
• Large Language Models
• Ollama Platform

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Machine Learning Engineer, Software Engineer, AI Engineer

Related on AIssential

• MiniMax M2.7 IS INSANE! Best Agentic/Coding Model! Beats Opus 4.6 and 50x Cheaper! (Fully Tested) — MiniMax M2.7 is an autonomously self-improving agentic AI model offering high performance at significantly lower costs.
• 5 Cool Things I Did with Local Language Models — Local LLMs offer superior privacy, cost, and offline capabilities for sensitive data and proprietary workflows.
• MiniMax Releases M2.1: An Enhanced M2 Version with Features like Multi-Coding Language Support, API Integration, and Improved Tools for Structured Coding — MiniMax M2.1 enhances open-source coding and automation with multi-language support and cost-efficiency.
• MiniMax M2.7 Goes Open-Weight to Let You Run Agents Locally — MiniMax M2.7's open-weight release enables local deployment for complex agentic and software engineering tasks.
• MiniMax Releases MiniMax M3 with MSA Architecture Supporting 1M-Token Context, Native Multimodality, and Agentic Coding — MiniMax M3's MSA architecture dramatically improves long-context efficiency and enables native multimodality and advanced agentic coding.
• GLM 5.2: the top Frontend Coding model in the world, IndexShare reduces costs — GLM-5.2 sets a new open-weight standard for long-horizon coding and agentic AI through architectural and inference optimizations.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Ollama Blog.