RTX 4090 × Meta Llama 3.1 8B Instruct
Meta Llama 3.1 8B Instruct on RTX 4090
125best gen tok/s
7,871best prompt tok/s
4,352max context tested
10results
Measured · all numbers below come from the linked sources
llama.cpp
| Quant | Gen tok/s | Prompt tok/s | Source | Context tested | TTFT | Build |
|---|---|---|---|---|---|---|
| Q4_K - Medium | 125 | 7,871 | LocalScore ↗ | 512+128 | 170 ms | llamafile 0.9.2 |
| Q4_K - Medium | 122 | 6,657 | LocalScore ↗ | 512+128 | 190 ms | llamafile 0.9.2 |
| Q4_K - Medium | 120 | 7,594 | LocalScore ↗ | 512+128 | 176 ms | llamafile 0.9.2 |
| Q4_K - Medium | 115 | 6,909 | LocalScore ↗ | 512+128 | 188 ms | llamafile 0.9.2 |
| Q4_K - Medium | 103 | 7,398 | LocalScore ↗ | 512+128 | 180 ms | llamafile 0.9.2 |
| Q4_K - Medium | 98.5 | 7,159 | LocalScore ↗ | 4,352 | 182 ms | llamafile 0.9.3 |
| Q4_K - Medium | 84.2 | 7,015 | LocalScore ↗ | 512+128 | 189 ms | llamafile 0.9.2 |
| Q4_K - Medium | 80.5 | 6,998 | LocalScore ↗ | 512+128 | 188 ms | llamafile 0.9.3 |
| Q4_K - Medium | 58.3 | 5,925 | LocalScore ↗ | 512+128 | 229 ms | llamafile 0.9.2 |
| Q4_K - Medium | 25.8 | 4,014 | LocalScore ↗ | 512+128 | 330 ms | llamafile 0.9.2 |
Ollama
No measured Ollama results yet. Submit yours →
MLX
No measured MLX results yet. Submit yours →
"Context tested" is the largest prompt+generation length actually benchmarked (tokens), not a memory-fit estimate.
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.