DEVOURED
MiniMax M3
MiniMax has released M3, an open-weights model achieving frontier-level performance in coding and agentic tasks, supporting multimodal input and 1 million token context windows.
What: MiniMax's M3 is an open-weights, multimodal model (supporting image and video input) that excels in coding and agentic work, scoring 59.0% on SWE-Bench Pro and 66.0% on Terminal Bench 2.1. It features a new Sparse Attention architecture, allowing context windows of up to 1 million tokens.
Why it matters: The release of an open-weights model with such advanced multimodal capabilities and an extremely long context window pushes the boundaries of what's publicly available, potentially accelerating innovation and accessibility for developers building agentic AI applications.
Takeaway: Developers interested in cutting-edge open-source models for coding or agentic applications should explore MiniMax M3 via MiniMax Code, Token Plan, or API services.
Decoder
- Open weights model: An AI model where the parameters (weights) are publicly released, allowing anyone to download, run, modify, and fine-tune it.
Original article
Introducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities
- Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas
- MiniMax Sparse Attention scales context to 1M
- Natively Multimodal from Step Zero
API: platform.minimax.io
Token Plan: platform.minimax.io/subscribe/toke…
🚀New! MiniMax Code: code.minimax.io
Weights & Tech Report in ~10 Days
API Pricing & Promotion
- 50% off standard usage (≤512K context) during the first 7 days
- Priority access available through: api@minimax.io
- Self-serve access for all users coming in the next few days
Explore all 68 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi