Search papers, labs, and topics across Lattice.
The paper introduces Nanbeige4.1-3B, a 3B parameter language model designed for strong agentic behavior, code generation, and general reasoning. It combines pointwise and pairwise reward modeling for preference alignment and complexity-aware rewards in reinforcement learning for code generation. The model is trained with complex data synthesis and turn-level supervision, enabling stable long-horizon tool interactions up to 600 turns.
A 3B parameter model now rivals models 10x its size in reasoning, alignment, and agentic tasks, challenging the assumption that bigger is always better.
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.
