VibeThinker-1.5B is an open-source reasoning model developed by WeiboAI (Sina Artificial Intelligence Lab), featuring 1.5 billion parameters. Despite its compact size, it demonstrates outstanding performance in mathematical and code reasoning tasks, surpassing some ultra-large-scale models like DeepSeek-R1.
Built on the innovative Spectrum-to-Signal Principle (SSP) training framework, VibeThinker achieves reasoning capabilities comparable to models with tens of billions of parameters, all while maintaining remarkably low training costs.
With only 1.5B parameters, VibeThinker demonstrates exceptional mathematical and logical reasoning capabilities, proving that intelligence doesn't always require massive scale.
Employs the Spectrum-to-Signal Principle (SSP), combining SFT (Spectrum Phase) and RL (Signal Phase) for two-stage optimization, enabling diverse exploration and precise refinement.
Total training time of approximately 3,900 GPU hours (around $7,800 USD), delivering exceptional performance-to-cost ratio compared to larger models.
Released under MIT open-source license, freely available for fine-tuning and commercial deployment, fostering innovation and research.
Designed to run efficiently on resource-constrained edge devices and research environments, making advanced AI accessible to more users.
Created by Sina's Artificial Intelligence Lab (WeiboAI), leveraging years of expertise in natural language processing and machine learning research.
VibeThinker's core innovation lies in its unique two-stage training methodology
Diversity & Exploration
Generates diverse reasoning paths and solution approaches, encouraging the model to explore "multiple possibilities" rather than converging prematurely on a single answer.
Refinement & Optimization
Through MaxEnt-Guided Policy Optimization (MGPO), reinforces correct signals and focuses optimization on high-uncertainty samples, refining the model's reasoning accuracy.
This "Diversity + Refinement" mechanism enables the model to think like humans: first diverge, then converge—exploring multiple approaches before settling on the optimal solution.
VibeThinker-1.5B demonstrates competitive performance across multiple reasoning benchmarks
| Benchmark | Score | Comparison |
|---|---|---|
| AIME24 | 80.3 | Leading among similar-sized models |
| AIME25 | 74.4 | Approaching large model performance |
| HMMT25 | 50.4 | Strong mathematical logic capability |
| LiveCodeBench v6 | 51.1 | Outperforms some 10B+ models |
VibeThinker-1.5B outperforms DeepSeek-R1 (671B parameters) in reasoning tasks, showcasing the potential of "small models, big intelligence."
Performance on general knowledge Q&A and encyclopedic tasks remains slightly behind ultra-large models, as expected for a specialized reasoning model.
Automated math tutoring systems, competition problem solvers, and educational platforms
Code generation, debugging assistance, and automated programming instruction
Algorithm analysis, symbolic reasoning, and computational research assistance
Lightweight local AI deployment on resource-constrained devices
Competition preparation, homework assistance, and interactive learning systems
Automated theorem proving, logical inference, and decision support systems
VibeThinker-1.5B represents a groundbreaking approach: "Use intelligent training methods, not blind parameter stacking." In an era of model miniaturization, reasoning enhancement, and efficient training, it is becoming an important milestone for next-generation high-performance open-source models.
Developed by WeiboAI (Sina Artificial Intelligence Lab), VibeThinker demonstrates that exceptional reasoning capabilities can be achieved with compact models through innovative training methodologies. This opens new possibilities for deploying advanced AI in resource-constrained environments and democratizing access to powerful reasoning models.
Key Takeaway: VibeThinker proves that with the right training approach, small models can achieve reasoning performance comparable to—or even exceeding—models hundreds of times larger, all while maintaining practical deployment costs and accessibility.