community pool / model answer
Qwen3-30B-A3B
chat · Qwen/Qwen3-30B-A3B
90.5 tok/s · model median · 8 runs in the model record.
| Reference rig | Quantization | Median result | Basis | TTFT (ms) | Peak VRAM | Max context | Engines |
|---|---|---|---|---|---|---|---|
| M1 Max 64GB | 4-bit | 62.6 tok/s | reported | 24 | - GB | 2,048 | ollama |
| M3 Ultra 512GB | 4-bit | 102 tok/s | reported | 76 | - GB | 8,192 | llama.cpp |
| RTX 2080 8GB | 4-bit | 19 tok/s | reported | - | 12 GB | 2,048 | ollama |
| RTX 3060 12GB | 4-bit | 107 tok/s | measured | - | 12 GB | 2,048 | llama.cpp, ollama |
| RTX 3080 12GB | 4-bit | 31 tok/s | reported | - | 8.2 GB | 2,048 | ollama |
| RTX3060+RTX2080-pooled-20GB 20GB | 4-bit | 91 tok/s | reported | - | 12 GB | 2,048 | ollama |
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.