dots3-note-prev model released: SWE-bench 78.4, designed for local use

victormustar · x · 2026-08-16

dots-studio released the dots3-note-prev model (280B/16B active) on Hugging Face, with benchmark scores: SWE-bench Verified 78.4, Terminal-Bench 2.1 75.1, ARC-AGI-2 81.4, ARC-AGI-3 general 32.1, LiveCodeBench v6 91.5, IFEval 93.9, SWE-bench-pro 61.0. The author expressed excitement to try it locally.

Related event: llama.cpp adds Dots3 Note model with SWE-bench 78.4 score(2 posts)→

Original post →

More from Models

Models channel →