SenseTime open-sources SenseNova-Loop-9B, a 9B looped-architecture model rivaling 27B models
pbaylies · x · 2026-11-17
SenseTime officially open-sourced SenseNova-Loop-9B (Apache-2.0), targeting deeper computation within a compact parameter budget.
- Architecture: reuses layer weights via a looped design combined with manifold-constrained hyper-connections (mHC) — 32 unique layers, 56 layer executions — aimed at complex reasoning, coding, tool use and multimodal tasks
- Benchmarks: big gains over Preview — NL2Repo 19.91→41.36, AutomationBench 14.33→36.67, GAIA 37.86→69.90; claimed competitive with Qwen3.8-27B across multiple benchmarks
- Deployment: Day-0 4-bit AWQ weights (4.9GB), VRAM as low as 9 GiB at 64K context
- Capabilities: up to 512K context with text and image inputs
- Weights and code are on GitHub and Hugging Face
Related event: SenseTime Open-Sources SenseNova-Loop-9B, Punching at 27B-Class Models(4 posts)→
More from Models
- Benchmark hacking? Labs' own SWE-bench scores beat independent runs by 4-5 points on average — RexDouglass · 2026-11-18
- Paying users complain: model tuned for 980M free users leaves subscribers with 'scraps' — Yamapama · 2026-11-18
- Users rip ChatGPT voice mode: random interruptions, accent drift it can't fix — Yamapama · 2026-11-18
- Oru'el claims Kanook beats NVIDIA Flora and OpenAI on RoboRewardBench by 30-50% — AashaySachdeva · 2026-11-18
- Flash model size chart puts Qwen3.8-Flash-Next smallest at 176B total parameters — BerryBlush_Rogers691 · 2026-11-18
- Agent-tuned MiMo scores 0/10 in a real loop while base Qwen 3.5 gets 7-8/10 — GapNew4766 · 2026-11-18