Hardware /
AMD Radeon RX 6650 XT
12measured results
109best 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 | 109 | 3,031 | LocalScore ↗ | 512+128 | llamafile 0.9.1 | Details → |
| Llama 3.2 1B Instruct | Q4_K - Medium | 84.0 | 2,494 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Llama 3.2 1B Instruct | Q4_K - Medium | 83.9 | 2,606 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Llama 3.2 1B Instruct | Q4_K - Medium | 83.0 | 2,592 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Llama 3.2 1B Instruct | Q4_K - Medium | 82.3 | 2,475 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Llama 3.2 1B Instruct | Q4_K - Medium | 82.2 | 2,600 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Gemma 3 1b | Q4_0 | 73.9 | 4,183 | LocalScore ↗ | 512+128 | llamafile 0.9.3 | Details → |
| Llama 2 7B | Q4_0 | 64.5 | 1,089 | llama.cpp (Vulkan) ↗ | 512+128 | llama.cpp Vulkan @dbb852b | Details → |
| Llama 2 7B | Q4_0 | 62.1 | 1,030 | llama.cpp (Vulkan) ↗ | 512+128 | llama.cpp Vulkan @dbb852b | Details → |
| Gemma 3 1b It | Q4_K - Medium | 60.7 | 3,528 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Qwen2.5.1 Coder 7B Instruct | Q4_K - Medium | 26.8 | 566 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Meta Llama 3.1 8B Instruct | Q4_K - Medium | 25.9 | 563 | LocalScore ↗ | 512+128 | llamafile 0.9.1 | 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.