Xiaomi's MiMo sparsity upgrade seen as mirroring DeepSeek's NSA-to-DSA move
teortaxesTex · x · 2026-09-24
Observers note Xiaomi's MiMo models are skipping a generation: V2.6 still uses the 2025-style MiMo Hybrid-SWA design, while HySparse—like NSA and MoBA—is a published research idea turned into engineering. The key change in HySparse2 is moving from block sparsity to token-level sparsity, seen as exactly DeepSeek's move from NSA to DSA (token top-k) in V3.2-Exp. HySparseV2 reads as a more cautious cousin of a future V4.1, keeping YOCO and full-attention backbone layers while gaining prefill and memory efficiency. Third-party analysis, not official.
Related event: Xiaomi's MiMo-V3 to Adopt New HySparse2 Architecture(3 posts)→
More from Models
- GPT-6 Astra only model to finish DrivingBench: 134.7m in 5:22 for $7.74 — petrusenko_max · 2026-09-24
- Opus 5.5 usage quota holds up well, user says it never ran out before reset — dotey · 2026-09-24
- Why GPT-6's rumored recurrent-depth architecture could change inference economics — panic_in_the_galaxy · 2026-09-24
- Jev, a No-Text Model Claiming 200x Faster Decisions, Sparks 'Is a Well-Formatted Wrong Answer Still a Hallucination?' — jamesbrooksco · 2026-09-24
- Hands-on: Jev Beats Gemini Flash on Accuracy, Latency, and Cost — cantrell · 2026-09-24
- User cancels switch to Codex after Opus 5.5's 40% price cut wins him back — ayushtweetshere · 2026-09-24