community-verified · measured pool data
Can the RTX PRO 6000 Blackwell 96GB ×4 run GLM-5.2-MXFP8-NVFP4-NF3-Hybrid?
Yes — community-measured at 74.4 tok/s (2 runs at 4-bit).
Verified single-stream runs on the RTX PRO 6000 Blackwell 96GB ×4 from the localmaxxing public pool. Every number on this page is measured — nothing here is estimated.
| Quantization | Median tok/s | TTFT (ms) | Peak VRAM | Max context tested | Engines |
|---|---|---|---|---|---|
| 4-bit | 74.4 | 175 | — | 480,000 | vllm |
| 8-bit | 68.4 | 282.8 | — | 480,000 | vllm |