SenseNova Loop 9B Claimed to Match Qwen3 27B at 9B Params, GGUF Coming
MeinDruckerSpinnt · reddit · 2026-11-17
A Redditor (with the authors' permission) previews SenseNova Loop 9B: a 9B looped model claimed to rival Qwen3 27B-class models if the benchmarks hold. The authors say a GGUF release lands tomorrow; vision is supported, and 256k context needs roughly 15GB VRAM. The poster found the preview version fast enough in testing weeks ago. Claims are author-reported and unverified by third parties.
More from Models
- Claude Code Power User Burns 53% of Weekly Limit in Two Days, Downgrades to Sonnet — Inevitable_Truth_85 · 2026-11-17
- SenseNova-Loop-9B Lands on Hugging Face, Claiming 27B-Class Performance — Marcuss2 · 2026-11-17
- AA-PersonalityBench ranks Gemini 4 Argon most metaphorical model, GPT-3.5 Turbo least — ArtificialAnlys · 2026-11-17
- Counterfactual benchmark shows 11 frontier models fail to grasp programming semantics — jessyjli · 2026-11-17
- Harness Adapters: RL-trained LoRA lifts Qwen3.8-27B on Pi from 29% to 63% on SWE-bench Pro — LysandreJik · 2026-11-17
- Qwen says offensive-cyber training data was removed — placebo or real lobotomy? — Rare_Subject2343 · 2026-11-17