community pool / model answer
GLM-4.5-Air
chat · zai-org/GLM-4.5-Air
103 tok/s · model median · 3 runs in the model record.
| Reference rig | Quantization | Median result | Basis | TTFT (ms) | Peak VRAM | Max context | Engines |
|---|---|---|---|---|---|---|---|
| RTX PRO 6000 Blackwell 96GB | 4-bit | 108.9 tok/s | reported | 32 | - GB | 32,768 | vllm |
| Ryzen AI Max 395 128GB | 4-bit | 6.2 tok/s | reported | 1,131 | - GB | 783 | vllm |
| Ryzen AI Max 395 64GB | 3-bit | 29.6 tok/s | reported | 3,742 | 49.9 GB | 8,192 | 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.