Sina has released VibeThinker-3B, a Chinese language model with three billion parameters that sometimes matches models a hundred times larger on math and coding tasks. According to a technical report, it performs on par with DeepSeek V3.2 and Kimi K2.5 on competitive benchmarks like AIME26, with those models having 200 to 333 times more parameters. Its predecessor, VibeThinker-1.5B, launched in November 2025. Across six math and coding benchmarks, the 3B model falls within the performance range of five current top models including Gemini 3 Pro, GLM-5, and Claude Opus 4.5. On structured tasks like math olympiads or programming challenges, VibeThinker-3B matches models like GLM-5 or Gemini 3 Pro, and on LiveCodeBench, it beats every other model under 20 billion parameters. However, on the knowledge-heavy GPQA-Diamond benchmark, the model falls well behind its much larger competitors. To rule out data contamination, the team had the model compete in LeetCode contests held between late April and late May 2026, after training wrapped up, where it solved 123 out of 128 problems on the first try, ahead of GPT-5.2, Qwen3-Max, Kimi K2.5, and Claude Opus 4.6. VibeThinker-3B builds on Al