Llama 2 7B on RTX 5060 Ti
97.0best gen tok/s
4,196best prompt tok/s
640max context tested
4results
Measured · all numbers below come from the linked sources
llama.cpp
| Quant | Gen tok/s | Prompt tok/s | Source | Context tested | TTFT | Build |
|---|---|---|---|---|---|---|
| Q4_0 | 97.0 | 3,913 | llama.cpp (Vulkan) ↗ | 512+128 | — | llama.cpp Vulkan @89f10ba — coopmat2 |
| Q4_0 | 93.5 | 3,461 | llama.cpp (Vulkan) ↗ | 512+128 | — | llama.cpp Vulkan @89f10ba — coopmat2 |
| Q4_0 | 93.5 | 4,196 | llama.cpp (CUDA) ↗ | 512+128 | — | llama.cpp CUDA @89d1029 |
| Q4_0 | 90.9 | 3,737 | llama.cpp (CUDA) ↗ | 512+128 | — | llama.cpp CUDA @89d1029 |
Ollama
No measured Ollama results yet. Submit yours →
MLX
No measured MLX results yet. Submit yours →
"Context tested" is the largest prompt+generation length actually benchmarked (tokens), not a memory-fit estimate.