Unsloth releases quantized MiniMax-M3.1-Flash GGUF; the 2-bit version can run locally with 150GB of memory
MiniMax-M3.1-Flash can now run locally!
AI BriefingUnsloth has released a quantized GGUF for MiniMax-M3.1-Flash. The 2-bit version retains 91.5% top-1 accuracy while reducing size by 45% and can run locally with 150GB of memory; the 1-bit version achieves about 89.9%, and 4-bit UD-Q4_K_XL reaches 98.0%. The author calls it one of the strongest open-source models at its size and provides a Hugging Face weights link and local setup guide.
Also reported by 1 other source
- MiniMax M3.1 Flash Can Run Locally with 143.9GB of Memory Using Unsloth Dynamic GGUFMiniMax · 21 days ago
Source: Unsloth · x.com
Event follow-up3
View all event coverage- Unsloth releases quantized MiniMax-M3.1-Flash GGUF; the 2-bit version can run locally with 150GB of memoryUnsloth · 21 days ago
- MiniMax releases M3.1 Flash weights: 428B total parameters, 23B active per tokenMiniMax · 25 days ago
- MiniMax released M3.1 Flash, nearly tripling inference speed and reducing usage costs by 70%MiniMax:Blog · 28 days ago