Hardware /
AMD Radeon RX 6600 XT
8measured results
53.9best 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 2 7B | Q4_0 | 53.9 | 575 | llama.cpp (Vulkan) ↗ | 512+128 | llama.cpp Vulkan @091592d | Details → |
| Phi 4 Mini Instruct | Q6_K | 28.2 | 806 | LocalScore ↗ | 4,352 | llamafile 0.9.3 | Details → |
| Qwen3 4B Instruct 2507 | Q4_K - Medium | 26.4 | 671 | LocalScore ↗ | 4,352 | llamafile 0.9.3 | Details → |
| Gemma 3 Finetune | Q8_0 | 24.6 | 1,181 | LocalScore ↗ | 4,352 | llamafile 0.9.3 | Details → |
| Qwen2.5.1 Coder 7B Instruct | Q6_K | 23.1 | 507 | LocalScore ↗ | 4,352 | llamafile 0.9.3 | Details → |
| Llama 3.1 8B Instruct Abliterated_Via_Adapter | Q4_K - Medium | 21.9 | 462 | LocalScore ↗ | 4,352 | llamafile 0.9.3 | Details → |
| Meta Llama 3.1 8B Instruct | Q4_K - Medium | 21.5 | 604 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Deepseek-R1-0528-Qwen3-8B | Q6_K | 18.3 | 432 | LocalScore ↗ | 4,352 | 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.