MiniMax-M1 brings 1M-token open reasoning model
MiniMax released M1, an open-source reasoning model with 1M-token context, 80k output, and low-cost API pricing.
Timeline
oMLX 0.4.5.dev1 adds custom kernels for GLM-5.2 and MiniMax M3, plus cache fixes and better model profile exposure.
MiniMax M3 brings 229.9B MoE weights, 1M context, and multimodal output, but it needs serious GPU memory to run.
MiniMax M3 brings coding and agent features, a 1 million-token context window, and multimodal input to the company’s flagship model.
MiniMax M3 makes a strong case that open-weight models can still lead on coding, context, and price.
MiniMax M3 is a real step forward because it pairs long context with multimodal and agentic control.
MiniMax M2.7 is a practical agent model for coding, office edits, and complex tasks, with a copy-ready API setup.
MiniMax M2.1 pushes multi-language coding, agent tools, and UI-heavy app work into one model you can actually use.
MiniMax Speech 2.6 cuts latency under 250 ms, reads messy formats better, and clones voices more fluently across 40+ languages.
MiniMax M2.7 matters because it turns model improvement into an agentic workflow, not just a benchmark race.
MiniMax open-sourced M2, a model for agents and code that costs $0.30 per million input tokens and is free for a limited time.
MiniMax released M1, an open-source reasoning model with 1M-token context, 80k output, and low-cost API pricing.