QwQ-32B
Alibaba's 32B open reasoning model that rivals DeepSeek-R1 at a fraction of the size.
About
QwQ-32B is Alibaba's Qwen-team reasoning model, fully released March 5, 2025 after a late-2024 preview. It applies scaled reinforcement learning on top of the Qwen2.5-32B base, with a 131K context window, and Alibaba reports it matching DeepSeek-R1 (671B) on key reasoning benchmarks: 79.5 on AIME24, 63.4 on LiveCodeBench, 73.1 on LiveBench, and 83.9 on IFEval.
It is Apache 2.0 on Hugging Face and ModelScope, and it runs on consumer-grade hardware, which made it one of the most downloaded open reasoning models of 2025. These are Alibaba's own benchmark figures, and like all RL-trained reasoning models it can be verbose and occasionally overthink simple questions.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
- AIME24: 79.5 (Alibaba-reported)
- LiveCodeBench: 63.4 (Alibaba-reported)
- LiveBench: 73.1 (Alibaba-reported)
- IFEval: 83.9 (Alibaba-reported)
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 9/10
Scored 9/10 (verified tier, metrics-v1). Evidence: One of the most downloaded open reasoning models of 2025; topped HF open-source rankings at release; Apache 2.0 on Hugging Face and ModelScope; Alibaba-reported benchmarks (AIME24 79.5, LiveCodeBench 63.4) rival DeepSeek-R1; widely reproduced by community. License: Apache 2.0.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 9.0/10
- Maintenance · recency, releases, issue load
- 8.0/10
- Openness · license + docs
- 10.0/10
- Manual review · sanity read + red-flag check
- 8.0/10
- · One of the most downloaded open reasoning models of 2025; topped HF open-source rankings at release
- · Apache 2.0 on Hugging Face and ModelScope
- · Alibaba-reported benchmarks (AIME24 79.5, LiveCodeBench 63.4) rival DeepSeek-R1; widely reproduced by community