community pool / model answer
Qwen3-32B
chat · Qwen/Qwen3-32B
22.8 tok/s · model median · 4 runs in the model record.
| Reference rig | Quantization | Median result | Basis | TTFT (ms) | Peak VRAM | Max context | Engines |
|---|---|---|---|---|---|---|---|
| M1 Max 64GB | 4-bit | 9 tok/s | reported | 114 | - GB | 2,048 | ollama |
| Radeon AI Pro R9700 32GB ×2 | 8-bit | 22.9 tok/s | reported | 71 | - GB | 32,768 | vllm |
| Radeon AI Pro R9700 32GB ×3 | 8-bit | 22.8 tok/s | reported | 60 | 34.2 GB | 786 | vllm |
| Radeon AI Pro R9700 57GB | 4-bit | 3.2 tok/s | reported | 8,835 | - GB | 2,048 | llama.cpp |
Context tested
The table above is the tested context for each exact rig and quantization cell. Empty fields remain empty rather than being filled with an estimate.