RTX 5060 Ti × Qwen2.5 14B Instruct
Qwen2.5 14B Instruct on RTX 5060 Ti
39.4best gen tok/s
1,493best prompt tok/s
4,352max context tested
14results
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 | 39.4 | 1,493 | LocalScore ↗ | 4,352 | 899 ms | llamafile 0.9.2 |
| Q4_K - Medium | 36.5 | 1,465 | LocalScore ↗ | 512+128 | 917 ms | llamafile 0.9.2 |
| Q4_K - Medium | 36.4 | 1,366 | LocalScore ↗ | 4,352 | 1,202 ms | llamafile 0.9.2 |
| Q4_K - Medium | 35.0 | 1,410 | LocalScore ↗ | 512+128 | 937 ms | llamafile 0.9.2 |
| Q4_K - Medium | 34.3 | 1,320 | LocalScore ↗ | 512+128 | 1,012 ms | llamafile 0.9.2 |
| Q4_K - Medium | 34.2 | 1,410 | LocalScore ↗ | 512+128 | 950 ms | llamafile 0.9.2 |
| Q4_K - Medium | 34.0 | 1,429 | LocalScore ↗ | 512+128 | 933 ms | llamafile 0.9.2 |
| Q4_K - Medium | 33.8 | 1,418 | LocalScore ↗ | 512+128 | 945 ms | llamafile 0.9.2 |
| Q4_K - Medium | 32.9 | 1,419 | LocalScore ↗ | 512+128 | 944 ms | llamafile 0.9.2 |
| Q4_K - Medium | 31.3 | 1,287 | LocalScore ↗ | 512+128 | 1,352 ms | llamafile 0.9.2 |
| Q4_K - Medium | 30.5 | 1,161 | LocalScore ↗ | 512+128 | 1,666 ms | llamafile 0.9.2 |
| Q4_K - Medium | 30.5 | 1,245 | LocalScore ↗ | 512+128 | 3,051 ms | llamafile 0.9.3 |
| Q4_K - Medium | 22.6 | 1,284 | LocalScore ↗ | 512+128 | 1,034 ms | llamafile 0.9.2 |
| Q4_K - Medium | 16.6 | 953 | LocalScore ↗ | 512+128 | 1,440 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.
What else can RTX 5060 Ti 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.