MiniMaxAI
MiniMax M3
Native multimodal MoE model with about 428B total and 23B activated parameters and 1M-token context, aimed at coding and agent work.MiniMax Community LicenseHugging Face
No server in stock right now has the 1025 GB this model needs for inference.
License restrictions
Commercial use requires a 'Built with MiniMax M3' notice and, above USD 20M yearly revenue, prior written authorization from MiniMax.
GPU memory it needs
From model cards and vendor documentation. Lower precision needs less memory, with slightly lower quality. Estimates are marked.