RTX A4000 × Meta Llama 3.1 8B Instruct
Meta Llama 3.1 8B Instruct on RTX A4000
60.6best gen tok/s
2,011best 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 | 60.6 | 2,011 | LocalScore ↗ | 512+128 | 656 ms | llamafile 0.9.2 |
| Q4_K - Medium | 60.0 | 1,989 | LocalScore ↗ | 512+128 | 654 ms | llamafile 0.9.2 |
| Q4_K - Medium | 59.1 | 1,949 | LocalScore ↗ | 4,352 | 674 ms | llamafile 0.9.2 |
| Q4_K - Medium | 59.0 | 1,963 | LocalScore ↗ | 512+128 | 671 ms | llamafile 0.9.2 |
| Q4_K - Medium | 58.9 | 1,945 | LocalScore ↗ | 4,352 | 675 ms | llamafile 0.9.2 |
| Q4_K - Medium | 58.8 | 1,932 | LocalScore ↗ | 4,352 | 679 ms | llamafile 0.9.2 |
| Q4_K - Medium | 58.6 | 1,843 | LocalScore ↗ | 4,352 | 686 ms | llamafile 0.9.2 |
| Q4_K - Medium | 58.5 | 1,910 | LocalScore ↗ | 4,352 | 685 ms | llamafile 0.9.2 |
| Q4_K - Medium | 46.0 | 1,885 | LocalScore ↗ | 512+128 | 686 ms | llamafile 0.9.3 |
| Q4_K - Medium | 45.4 | 1,881 | LocalScore ↗ | 512+128 | 686 ms | llamafile 0.9.3 |
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.
What else can RTX A4000 run? →
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.