Nanbeige4.3-33B: open 33B model claims top agentic scores in its class
Why it matters
Nanbeige LLM Lab released Nanbeige4.3-33B on Hugging Face (Oct 24, 2026), an open-weights model under Apache 2.0 that its model card calls the flagship teacher model of the Nanbeige4.3 series, delivering state-of-the-art agentic performance among open models in the ~30B class. On the lab's own benchmarks (all models at highest thinking effort, same harness), it scores 80.9 on Terminal-Bench 2.1 vs 73.0 for Qwen3.8-27B, 61.8 for Muse-Glimmer-30B and 47.2 for Gemma4-31B; 65.5 on DeepSWE vs 42.5 / 4.4 / 2.7; and 23.8 on Terminal-Bench 4.0 vs 6.3 / 1.6 / 0.0. It trails on a few: SWE-Bench Pro 60.3 vs 61.7 and GDPval 84.8 vs 86.4 (Qwen3.8-27B), GPQA Diamond 89.5 vs 91.3 (Gemma4-31B). Context goes up to 524,288 tokens (512K), and the lab plans to distill it into smaller models such as an upcoming Nanbeige4.3-3B. These are lab-reported numbers, not independent evaluations.
Read the source: Hugging Face (Nanbeige LLM Lab) (opens in a new tab)
All hits from 25 Oct 2026More Open ModelsGet the daily hits by email