Hardware /
Orin
4measured results
47.8best gen tok/s*
7 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 | 47.8 | 1,233 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Llama 3.2 1B Instruct | Q4_K - Medium | 42.7 | 1,082 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Llama 2 7B | Q4_0 | 35.9 | 1,172 | llama.cpp (CUDA) ↗ | 512+128 | llama.cpp CUDA @c1b1876 | Details → |
| Llama 2 7B | Q4_0 | 33.6 | 991 | llama.cpp (CUDA) ↗ | 512+128 | llama.cpp CUDA @c1b1876 | 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.