Nanbeige4.1-3B

Nanbeige4.1-3B is an enhanced iteration of the Nanbeige4-3B-Base model, achieved through further post-training optimization with supervised fine-tuning (SFT) and reinforcement learning (RL). This model demonstrates strong reasoning capabilities, robust preference alignment, and effective agentic behaviors, making it a competitive open-source model at a small parameter scale.

Nanbeige4.1-3B is an enhanced iteration of the Nanbeige4-3B-Base model, achieved through further post-training optimization with supervised fine-tuning (SFT) and reinforcement learning (RL). This model demonstrates strong reasoning capabilities, robust preference alignment, and effective agentic behaviors, making it a competitive open-source model at a small parameter scale.

• 01Strong reasoning capabilities
• 02Robust preference alignment
• 03Effective agentic behaviors
• 04Superior performance on general reasoning tasks
• 05Native support for deep-search tasks