Nanbeige4.1-3B: A Small General Model that Reasons, Aligns, and Acts

TL;DR

Nanbeige4.1-3B is a small, 3-billion-parameter generalist language model that excels in reasoning, code generation, and agentic behavior, outperforming larger models through innovative training techniques.

Contribution

The paper introduces Nanbeige4.1-3B, the first open-source small model to combine reasoning, code, and agentic capabilities in a single model with advanced training methods.

Findings

Outperforms similar-sized models like Nanbeige4-3B-2511 and Qwen3-4B.

Achieves superior performance compared to larger models such as Qwen3-30B.

Executes up to 600 tool-call turns reliably for complex tasks.

Abstract

We present Nanbeige4.1-3B, a unified generalist language model that simultaneously achieves strong agentic behavior, code generation, and general reasoning with only 3B parameters. To the best of our knowledge, it is the first open-source small language model (SLM) to achieve such versatility in a single model. To improve reasoning and preference alignment, we combine point-wise and pair-wise reward modeling, ensuring high-quality, human-aligned responses. For code generation, we design complexity-aware rewards in Reinforcement Learning, optimizing both correctness and efficiency. In deep search, we perform complex data synthesis and incorporate turn-level supervision during training. This enables stable long-horizon tool interactions, allowing Nanbeige4.1-3B to reliably execute up to 600 tool-call turns for complex problem-solving. Extensive experimental results show that Nanbeige4.1-3B significantly outperforms prior models of similar scale, such as Nanbeige4-3B-2511 and Qwen3-4B, even achieving superior performance compared to much larger models, such as Qwen3-30B-A3B. Our results demonstrate that small models can achieve both broad competence and strong specialization simultaneously, redefining the potential of 3B parameter models.