Hardware /
RTX A2000
5measured results
117best gen tok/s*
6 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 | 117 | 3,811 | LocalScore ↗ | 512+128 | llamafile 0.9.3 | Details → |
| Qwen3 Mini Moe | Q4_K - Medium | 85.1 | 5,172 | LocalScore ↗ | 512+128 | llamafile 0.9.3 | Details → |
| Llama 2 7B | Q4_0 | 45.7 | 1,362 | llama.cpp (Vulkan) ↗ | 512+128 | llama.cpp Vulkan @b1afcab — coopmat2 | Details → |
| Llama 2 7B | Q4_0 | 45.5 | 1,245 | llama.cpp (Vulkan) ↗ | 512+128 | llama.cpp Vulkan @b1afcab — coopmat2 | Details → |
| Meta Llama 3.1 8B Instruct | Q4_K - Medium | 28.7 | 758 | LocalScore ↗ | 512+128 | llamafile 0.9.3 | 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.