QwQ 32B
QwQ is the reasoning model of the Qwen series. Compared with conventional instruction-tuned models, QwQ, which is capable of thinking and reasoning, can achieve significantly enhanced performance in downstream tasks, especially hard problems.
Input
Output
Context
33K
Max Output
33K
Parameters
32B
Technical Specifications
Model TypeChat
Context Window33K
Max Output Tokens33K
Parameters32B
Training Cutoff—
LicenseQwen
Capabilities
Input Modalities
text
Output Modalities
text
Features
chain_of_thoughtreasoningstreamingtext_generation
Benchmarks
#179
Rank (Text)
+10/-10
arena text ci 95
1170
Arena Score (Text)
3,410
Votes (Text)
Resources & Links
Estimated GPU Requirements for QwQ 32B
Estimates based on INT8 quantization. Actual requirements vary by framework and configuration.
Related Tools
Data sourced from official provider APIs and documentation
Last updated: May 5, 2026
Start building with the right model.
From model selection to production, one platform, no fragmentation.