Nanbeige/Nanbeige4.3-33B
just landedmodel
We compare Nanbeige4.3-33B with recently released frontier models at around the 30B-parameter scale: Qwen3.8-27B, Muse-Glimmer-30B, and Gemma4-31B. Dense, 33B params. Context: 512K tokens (config). License: apache-2.0. Vendor-reported: "Delivering state-of-the-art agentic performance among open models in the ~30B parameter class, Nanbeige4.3-33B excels at complex coding and professional cowork tasks." Vendor-reported: "For example, on JobBench, it outperforms the baseline in most of the 35 occupations evaluated."
- Vendor-reported: Delivering state-of-the-art agentic performance among open models in the ~30B parameter class, Nanbeige4.3-33B excels at complex coding and professional cowork tasks.
- Vendor-reported: For example, on JobBench, it outperforms the baseline in most of the 35 occupations evaluated.
- Org
- Nanbeige
- Released
- 2026-10-24
- Status
- just landed
- License
- apache-2.0 (permissive)
- Parameters
- 33.1B
- Architecture
- dense · LlamaForCausalLM
- Context
- 524,288 tokens
- Task
- text-generation · text
- Signal
- 163 HF likes · 0 downloads
Runs on
mixed: real files + estimate; need = weights + FP16 KV cache for 8K tokens + 1GB. A sizing calculation, not tested.
- Q4 ~21.8 GB: 24GB GPU (3090 / 4090), 32GB GPU (5090), 24GB GPU (7900 XTX), Mac 36GB, Mac 48GB, CPU only, 32GB RAM (slow)
- Q8 ~37.9 GB (estimated size): 48GB GPU (RTX 6000 Ada / A6000), Mac 64GB (M1 Max), CPU only, 64GB RAM (slow)
- FP16 ~68.9 GB: 96GB GPU (RTX Pro 6000 Blackwell), 1× DGX Spark (128GB), 2× DGX Spark (256GB), Strix Halo 128GB unified (~96GB usable), Mac 96GB, Mac 128GB, Mac 192GB, Mac 256GB, Mac 512GB, CPU only, 128GB RAM (slow), CPU only, 256GB RAM (slow)
Links
First published here 2026-10-24 · updated 2026-10-25 · markdown · JSON