community pool / model answer
Muse-Glimmer-30B
chat · meta-models/Muse-Glimmer-30B
41.9 tok/s · model median · 17 runs in the model record.
| Reference rig | Quantization | Median result | Basis | TTFT (ms) | Peak VRAM | Max context | Engines |
|---|---|---|---|---|---|---|---|
| Arc Pro B70 32GB | 4-bit | 26.8 tok/s | measured | - | - GB | 131,072 | llama.cpp |
| Radeon AI Pro R9700 32GB ×3 | 4-bit | 52.8 tok/s | reported | 1,372 | - GB | 1,923 | llama.cpp |
| Radeon AI Pro R9700 32GB ×3 | 8-bit | 21.3 tok/s | measured | 875 | - GB | 1,923 | llama.cpp |
| RTX PRO 6000 Blackwell 96GB | 16-bit | 48.6 tok/s | reported | 239 | 52.3 GB | 131,072 | llama.cpp, vllm |
| Ryzen AI Max 395 64GB | 4-bit | 12.9 tok/s | reported | 4,689 | 16.4 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.