QwQ-32B
Open Weights
Alibaba Cloud
Alibaba's open-weight 32B reasoning model (March 2025, Apache 2.0), which uses reinforcement learning to reach reasoning performance competitive with much larger models and to beat o1 on some benchmarks, at a size that runs on a single high-end GPU.
Strengths
- Open-weight 32B reasoning model under Apache 2.0
- RL-trained to rival much larger reasoning models
- Beats o1 on some math and reasoning benchmarks
- 131K context window
- Runs on a single high-end consumer/prosumer GPU
Caveats
- Reasoning tuning can produce long, verbose chains
- Text-only (no vision or audio input)
- 32B ceiling below frontier reasoning models on the hardest tasks
- Vendor
- Alibaba Cloud
- Context Window
- 131,072 tokens
- License
- Apache-2.0
- Release Date
- 2025-03-06
- Modalities
- text
| Benchmark | Score | Max |
|---|---|---|
| livebench_instruction_following | 85.3 | 100.0 |
| livebench_language | 50.6 | 100.0 |
| livebench_coding | 72.2 | 100.0 |
Capabilities
Vision
Audio
Video
Tool Use
Resources
Pricing
Open weights under Apache 2.0 - free to download and self-host; 32B runs on a single high-end GPU.