Hugging Face 热门模型第 6 名：Nanbeige/Nanbeige4.3-33B，362 赞，近期下载 0 次。
标签：transformers, safetensors, llama, text-generation, llm, nanbeige, conversational, en, zh, base_model:Nanbeige/Nanbeige4.3-33B-Base, base_model:finetune:Nanbeige/Nanbeige4.3-33B-Base, license:apache-2.0
README：
1. Introduction
In Nanbeige4.3, we take a foundational step toward pushing the boundaries of lightweight agents. Our strategy centers on building a high-capacity teacher agent first, which will subsequently anchor the distillation of advanced agentic capabilities into smaller models.
To this end, we first introduce Nanbeige4.3-33B, designed as the flagship teacher model for the Nanbeige4.3 series. 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. It not only offers a powerful, cost-effective option for immediate real-world deployment, but also lays the necessary foundation to power our upcoming small model distillation pipeline (e.g., for Nanbeige4.3-3B).
2. Model Performance
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. All models are evaluated at their highest thinking effort. For each benchmark, we use the same inference hyperparameters and harness across models.
CategoryBenchmarkNanbeige4.3-33B1Qwen3.8-27BMuse-Glimmer-30BGemma4-31B
Agentic CodingTerminal-Bench 2.1280.973.061.847.2
Terminal-Bench 4.0323.86.31.60.0
DeepSWE365.542.54.42.7
SWE-Bench Pro360.361.736.951.2
SWE-Atlas-QnA450.841.917.73.3
NL2Repo553.048.327.433.8
Scicode52.646.744.945.5
SkillsBench61.156.439.728.9
Agentic CoworkAutomation Bench654.238.315.29.7
JobBench744.834.516.69.0
GDPval884.886.475.871.4
Claw-Evalpass^3978.375.257.344.2
Toolathlon-Verified1070.468.547.227.8
DeepPlanning1263.349.221.419.4
EnterpriseOps-Gym1241.040.230.527.3
YC-bench11$1,025,437$641,290$513,978$265,152
Tau3-Bench1248.848.533.718.6
ReasoningGPQA Diamond89.589.283.591.3
AIME 202695.294.594.789.2
AA-LCR84.082.083.372.0
1 Nanbeige4.3-33B is evaluated with preserve_thinking=true in the chat template.
2 Terminal-Bench 2.1 is evaluated using Terminus 2.
3 Terminal-Bench 4 and DeepSWE are evaluated using mini-swe-agent v2.4.6.
4 SWE-Atlas-QnA is evaluated using mini-swe-agent v2.4.6. Answers are evaluated against 1,304 task-specific rubrics using GPT-6 Sol as the judge, with medium reasoning effort.
5 NL2Repo is evaluated using Claude Code.
6 Automation Bench v1.0.6 includes 600 publicly released tasks.
7 JobBench is evaluated with the Claude Code harness equipped with a search tool.
8 GDPval is scored using revised rubrics.
9 Claw-Eval is evaluated exclusively on the Claw-Eval General subset (comprising 157 single-turn user interaction tasks).
10 Toolathlon-Verified sets no max_tokens limit for individual model requests.
11 YC-bench retains only the most recent 10 turns of user dialogue as historical context before each model call.
12 DeepPlanning, EnterpriseOps-Gym, Tau3-Bench are evaluated using their official repositories.
Nanbeige4.3-33B leads on most of the benchmarks, showing strong performance in both agentic coding and cowork tasks.
Beyond its breadth across benchmarks, Nanbeige4.3-33B shows strong performance on agent tasks across a wide range of professions and industries. For example, on JobBench, it outperforms the baseline in most of the 35 occupations evaluated.
3. Quickstart
The tokenizer provides a configurable chat template for reasoning and tool-use scenarios:
enable_thinking controls whether the mod
The original belongs to Hugging Face 热门模型 and its authors and is reproduced here for reading. To have it removed, write to support@aisa.one.
Studio