VibeThinker-3B: a 3B model matches or beats DeepSeek V3.2, GLM-5, and Gemini 3 Pro on verifiable reasoning

VibeThinker: 3B param model that beats Opus 4.5 on reasoning with novel SFT+GRPO

This tech report presents VibeThinker-3B, a 3B model that scores 94.3 on AIME26 (97.1 with test-time scaling), 80.2 Pass@1 on LiveCodeBench v6, and 96.1% on unseen LeetCode contests. It matches or exceeds DeepSeek V3.2, GLM-5, and Gemini 3 Pro on these verifiable tasks. The training pipeline uses curriculum-based SFT, multi-domain RL (GRPO), and offline self-distillation. IFEval stays at 93.4, so instruction following isn't sacrificed. The authors propose a Parametric Compression-Coverage Hypothesis: verifiable reasoning compresses into small cores, but open-domain knowledge still needs broad parameter coverage. The post doesn't disclose training data size, compute cost, or inference latency.

Why it matters: A 3B model matching or beating large models on math and coding reasoning is a strong contrast story with concrete numbers and a reproducible method. Deduction because it's a single paper not yet replicated by the community—82 feels right.