QwQ-32B
by Alibaba / Qwen Team · open-source · Last verified 2026-03-17
QwQ-32B is Alibaba's 32 billion parameter reasoning-focused language model that employs deep chain-of-thought reasoning to tackle complex mathematical, scientific, and logical problems. It achieves performance competitive with much larger models on reasoning benchmarks, demonstrating that focused reasoning training can be highly parameter-efficient.
https://huggingface.co/Qwen/QwQ-32B-Preview ↗B
B—Above Average
Adoption: BQuality: A+Freshness: ACitations: B+Engagement: F
Specifications
- License
- Apache 2.0
- Pricing
- open-source
- Capabilities
- mathematical-reasoning, logical-reasoning, chain-of-thought, code-reasoning, scientific-problem-solving
- Integrations
- Hugging Face, Ollama, vLLM
- Use Cases
- complex math problem solving, scientific research assistance, competitive programming, logical puzzle solving
- API Available
- Yes
- Parameters
- ~32B
- Context Window
- 32K
- Modalities
- text
- Training Cutoff
- 2024
- Tags
- reasoning, qwen, alibaba, chain-of-thought, math
- Added
- 2026-03-17
- Completeness
- 100%
Index Score
62Adoption
65
Quality
90
Freshness
82
Citations
72
Engagement
0