MiniMax open-sources M3.1 Flash: 428B MoE with 23B active, 1M context, live on Together AI
MiniMax_AI · x · 2026-10-23
- MiniMax announced that M3.1 Flash weights are now open, with Together AI as a day-0 launch partner offering fully managed serving.
- Native multimodal MoE: 428B total parameters with 23B active per token, accepts text, image, and video input, and supports up to a 1M-token context window.
- Positioned for agentic workflows, coding, computer use, and cybersecurity: it can reason over real codebases, fix bugs, build features, and turn screenshots/charts into polished frontends and presentations.
- Technical details: MiniMax Sparse Attention runs throughout the model, quantization-aware training covers routed experts and attention paths to cut memory footprint, and a DSpark speculative decoding head ships with the release.
- Released under the MiniMax Community License; Together offers it on serverless and dedicated infrastructure with a 99.9% SLA.
Related event: MiniMax Open-Sources M3.1-Flash: 428B-Parameter MoE with 1M Context(9 posts)→
More from coding & agent
- AI Exposes CVEs Faster Than Maintainers Can Patch: Dev Says AI Does More Harm Than Good — _jaydeepkarale · 2026-10-23
- AI is becoming a mathematical brainstorming companion for algorithm design — Fowe · 2026-10-23
- Peking University and Alibaba unveil OmniDR-Bench for video-centered deep research agents — Qwen · 2026-10-23
- OpenPrefs MCP centralizes your preferences across Claude, ChatGPT and agents — openprefs · 2026-10-23
- $100K Open Kernel Challenge accused of copying GPU MODE OSS code via AI agent rewrite — josh_wills · 2026-10-23
- GBrain Adds Multi-Account Switching So Your AI Memory Follows Each Identity — garrytan · 2026-10-23