Qwen3-32B
Qwen3-32B is an open-weights language model from Qwen, released April 29, 2025. It has 33B parameters (dense). The context window is 128K tokens. On the Epoch Capabilities Index it scores 136.9, rank 134 of 256 rated models. API price: $0.7 input / $2.8 output per million tokens.
Check it against your own PC →
Can you run Qwen3-32B locally?
The smallest common GPU it runs on is 16 GB of VRAM (with 32 GB of RAM): IQ3_XXS, split GPU / CPU, roughly 9 tokens per second. Estimates assume 32 GB of DDR5 RAM, a 16K context and llama.cpp-style GGUF quantization. Enter your exact hardware in the live app.
| GPU memory | Verdict | Best quant | VRAM | RAM | tok/s |
|---|---|---|---|---|---|
| 6 GB VRAM | Very slow | IQ3_XXS (noticeable loss) | 5.2 GB | 12.2 GB | ~2 |
| 8 GB VRAM | Very slow | IQ3_XXS (noticeable loss) | 7.2 GB | 10.2 GB | ~3 |
| 12 GB VRAM | Very slow | IQ3_XXS (noticeable loss) | 11.2 GB | 6.2 GB | ~4 |
| 16 GB VRAM | Runs, slowly | IQ3_XXS (noticeable loss) | 15.2 GB | 2.2 GB | ~9 |
| 24 GB VRAM | Runs on GPU | IQ4_XS (good) | 22.3 GB | – | ~31 |
| 32 GB VRAM | Runs on GPU | Q5_K_M (excellent) | 28.2 GB | – | ~44 |
Ready-made quantizations on Hugging Face: Qwen/Qwen3-32B-GGUF · unsloth/Qwen3-32B-GGUF · MaziyarPanahi/Qwen3-32B-GGUF · bartowski/Qwen_Qwen3-32B-GGUF · Qwen/Qwen3-32B-AWQ · RedHatAI/Qwen3-32B-NVFP4
Benchmark results
| Benchmark (Epoch AI) | Score |
|---|---|
| OTIS Mock AIME | 66.9% |
| GPQA Diamond | 56.6% |
| Aider Polyglot | 40.0% |
Links
Hugging Face: Qwen/Qwen3-32B · Epoch AI benchmarks
Models with a similar score
Gemini 2.5 Pro (May 2025) · DeepSeek-V3.1 · Gemini 2.5 Flash (Jun 2025) · Gemini 2.5 Flash (Apr 2025) · DeepSeek-R1 (May 2025) · o3-mini · gpt-oss-120b · Grok-3 mini
More from Qwen
Qwen3.8 Omni Flash · Qwen3.8 Max (0902) · Qwen3.8 Flash · Qwen3.8 Flash Next · Qwen 3.8 27B · Qwen3.8-2.4T-A95B · Qwen 3.8 Max · Qwen3.7 Flash