Nanbeige4.1-3B
Nanbeige4.1-3B is a 4B-parameter small language model that excels in reasoning, preference alignment, and agentic behaviors, proving that compact models can compete with much larger ones. Built on Nanbeige4-3B-Base, it achieves SOTA performance on benchmarks like LiveCodeBench-Pro (81.4% Easy, 28.1% Medium), AIME 2026 I (87.40%), Arena-Hard-v2 (73.2), and deep-search tasks (xBench-DeepSearch-2505: 75, xBench-DeepSearch-2510: 39). It is the first small model to natively support deep-search tasks with 500+ tool invocations, filling a gap in the small-model ecosystem. The model is open-source and deployable via transformers, with 131K max tokens and BF16 precision.
Transparency Score
Weighted across four pillars · updated July 1, 2026
Capabilities
Vendor Information
Complete information about the vendor/provider of this AI application
EU AI Act Provider Information
Potential Risks
1 considerations identified
Review recommended before use
These considerations are automatically identified based on publicly available information about the vendor and AI catalog data. Actual risks may vary based on your specific use case and implementation.
Get insights into risk by running assessments on this AI application.
Work at Nanbeige LLM lab? Claim this listing to correct or complete the data.
Ready to manage AI applications?
Track, assess, and govern your AI applications with Anove.