VibeThinker-3B: A 3B Dense Reasoning Model Built on Qwen2.5-Coder-3B With the Spectrum-to-Signal Post-Training Pipeline

VibeThinker-3B is a 3 billion parameter dense reasoning model developed on the Qwen2.5-Coder-3B architecture, utilizing the Spectrum-to-Signal post-training pipeline. It demonstrates competitive performance against models such as DeepSeek V3.2 and Kimi K2.5 on verifiable benchmarks. This model is significant for practitioners as it offers a robust option for reasoning tasks, enhancing capabilities in applications requiring dense reasoning.