@XiaomiMiMo is already working on next MiMo-V3 architecture!
This is a dev post classified by Jev as Other (a launch), kept by the Dev Radar because it carries real work, not commentary.
This is awesome! @XiaomiMiMo is already working on next MiMo-V3 architecture! So there might not be anymore v2 models, MiMo-v2.6 might be the last one. Their motivation is interesting. Their paper HySparse2: Hybrid Sparse Attention with Two-Level KV Sharing has it very clear. Reading through paper, what i learned is this will result in much improved - prompt processing speed - TTFT - prefill energy usage - GPU requirements for prefill servers Cause prompt doesn't necessarily need to travel through the entire network. I really like these kind of research going on to making inference effic
Posted by AJ (7.7k followers) 1 h ago · 8 likes · 627 views · view the original post on X. Kept by the Dev Radar as Other.
More dev work like this
- 🚨重磅！ MiMo-V2.6 刚登顶开源，小米AI负责人罗福莉直接预告 MiMo-V3换心脏！ — @NFT_Chen
- Run uncensored Bonsai 2 27B on 16 GB MacBook 💻 — @atomic_chat_hq
- Marking our third audio model release in < 30 days, today we're introducing two new… — @NewsFromGoogle
- Meet XYZ-Aquila-mini-APEX-I-MiniPlus-V2.1-GGUF: a quantized MoE model built for… — @HuggingModels
- Eikos, um modelo de decisão open source (MIT), alternativa ao Jev e ao Laya. — @0xCVYH
- The name HySparse2 really makes me think of Hy (Hunyuan). I even catch myself reading it… — @sheriyuo
- Not bad — @zephyr_z9
- Xiaomi is skipping a generation. V2.6 still uses MiMo Hybrid-SWA, a very 2025 design;… — @teortaxesTex
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 22.6k posts from 5k X accounts over the last 21 days, 2.6k tools, 12 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-23 15:57 UTC. Full method.