Unsloth squeezes 428B MiniMax M3.1 Flash into 143.9GB 2-bit GGUF with 91.5% accuracy
danielhanchen · x · 2026-10-27
MiniMax announced that Unsloth's Dynamic GGUFs bring its 428B MoE M3.1 Flash to local machines: a 2-bit quantized version at just 143.9GB (45% smaller), retaining 91.5% top-1 token agreement across seven domains, runnable on 150GB RAM via Unsloth Desktop or llama.cpp with an OpenAI-compatible API. Unsloth calls it the strongest open model for its size and among the best overall.
Related event: Unsloth's 2-bit Quantization Runs MiniMax M3.1 Flash Locally(2 posts)→
More from Infra
- OpenAI's initial Vera CPU results: 51% more Python throughput vs leading x86 — firstadopter · 2026-10-27
- Berth: Open-Source Tool Lets Multiple AI Engines Share One GPU Without Fighting — Distinct-Pie2389 · 2026-10-27
- Cloudflare discloses cross-request KV cache leak in GLM-5.3 on Workers AI: 155 responses affected — michellechen · 2026-10-27
- Cloudflare discloses cross-request KV cache exposure in GLM-5.3 on Workers AI (155 responses) — wondersky · 2026-10-27
- One month fully local on a DGX Spark: daily tasks fine, code still needs Claude to check — michaelthatsit · 2026-10-27
- Cloudflare Workers adds spanContext, setStatus and updateName tracing APIs — irvinebroque · 2026-10-27