Hardware /
RTX 2070 SUPER
8measured results
212best gen tok/s*
8 GBreported memory
*Headline excludes tiny/test models (under 0.6B params or TinyStories-style). Open in rig checker →
llama.cpp
| Model | Quant | Gen tok/s | Prompt tok/s | Source | Context tested | Build | |
|---|---|---|---|---|---|---|---|
| Llama 3.2 1B Instruct | Q4_K - Medium | 212 | 6,350 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Llama 3.2 1B Instruct | Q4_K - Medium | 212 | 6,082 | LocalScore ↗ | 4,352 | llamafile 0.9.2 | Details → |
| Llama 3.2 1B Instruct | Q4_K - Medium | 189 | 5,696 | LocalScore ↗ | 4,352 | llamafile 0.9.2 | Details → |
| Llama 2 7B | Q4_0 | 88.1 | 2,088 | llama.cpp (CUDA) ↗ | 512+128 | llama.cpp CUDA @bc07349 | Details → |
| Llama 2 7B | Q4_0 | 87.7 | 2,293 | llama.cpp (CUDA) ↗ | 512+128 | llama.cpp CUDA @bc07349 | Details → |
| Llama 2 7B | Q4_0 | 64.6 | 1,199 | llama.cpp (Vulkan) ↗ | 512+128 | llama.cpp Vulkan @b7552cf | Details → |
| Meta Llama 3.1 8B Instruct | Q4_K - Medium | 57.8 | 1,460 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Meta Llama 3.1 8B Instruct | Q4_K - Medium | 57.4 | 1,428 | LocalScore ↗ | 4,352 | llamafile 0.9.2 | Details → |
Ollama
No measured Ollama results yet. Submit yours →
MLX
No measured MLX results yet. Submit yours →
Includes results from LocalScore, a Mozilla Builders project by cjpais, mirrored in the cn0303/local-llm-speed-benchmarks dataset. LocalScore runs use llamafile, which is built on llama.cpp.