mamba2-2.7b
6 records across 2 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RTX PRO 6000 | mamba2-2.7b 2.7B | Q4_K_M | llama.cpp (CUDA) | 1638.9 | – | – | – | – | 2026-08 | community | – | source |
| RTX PRO 6000 | mamba2-2.7b 2.7B | BF16 | llama.cpp (CUDA) | 1508.0 | – | – | – | – | 2026-08 | community | – | source |
| RTX PRO 6000 | mamba2-2.7b 2.7B | Q4_K_M | llama.cpp (CUDA) | 953.0 | – | – | – | – | 2026-08 | community | – | source |
| RTX PRO 6000 | mamba2-2.7b 2.7B | BF16 | llama.cpp (CUDA) | 630.9 | – | – | – | – | 2026-08 | community | – | source |
| GB10 | mamba2-2.7b 2.7B | Q8_0 | llama.cpp (CUDA) | 246.7 | – | – | – | – | 2026-08 | community | – | source |
| GB10 | mamba2-2.7b 2.7B | Q8_0 | llama.cpp (CUDA) | 98.4 | – | – | – | – | 2026-08 | community | – | source |