RTX 4050 Laptop GPU × TinyStories 656K
TinyStories 656K on RTX 4050 Laptop GPU
1,905best gen tok/s
159,348best prompt tok/s
640max context tested
23results
Measured · all numbers below come from the linked sources
llama.cpp
23 tiny/test model result(s) (excluded from headlines)
| Quant | Gen tok/s | Prompt tok/s | Source | Context tested | TTFT | Build |
|---|---|---|---|---|---|---|
| Q4_K - Medium | 1,905 | 145,343 | LocalScore ↗ | 512+128 | 10 ms | llamafile 0.9.3 |
| Q4_K - Medium | 1,890 | 144,854 | LocalScore ↗ | 512+128 | 10 ms | llamafile 0.9.3 |
| Q5_1 | 1,890 | 145,629 | LocalScore ↗ | 512+128 | 11 ms | llamafile 0.9.3 |
| Q5_K - Small | 1,890 | 142,236 | LocalScore ↗ | 512+128 | 10 ms | llamafile 0.9.3 |
| Q4_1 | 1,887 | 148,849 | LocalScore ↗ | 512+128 | 10 ms | llamafile 0.9.3 |
| Q4_K - Small | 1,885 | 140,884 | LocalScore ↗ | 512+128 | 10 ms | llamafile 0.9.3 |
| Q5_K - Medium | 1,879 | 153,054 | LocalScore ↗ | 512+128 | 9 ms | llamafile 0.9.3 |
| Q6_K | 1,873 | 159,348 | LocalScore ↗ | 512+128 | 9 ms | llamafile 0.9.3 |
| Q5_K - Medium | 1,869 | 145,396 | LocalScore ↗ | 512+128 | 10 ms | llamafile 0.9.3 |
| Q4_0 | 1,859 | 147,634 | LocalScore ↗ | 512+128 | 10 ms | llamafile 0.9.3 |
| Q5_0 | 1,856 | 154,382 | LocalScore ↗ | 512+128 | 10 ms | llamafile 0.9.3 |
| Q4_K - Medium | 1,732 | 153,201 | LocalScore ↗ | 512+128 | 9 ms | llamafile 0.9.3 |
| Q3_K - Medium | 898 | 38,782 | LocalScore ↗ | 512+128 | 39 ms | llamafile 0.9.3 |
| Q3_K - Large | 865 | 39,062 | LocalScore ↗ | 512+128 | 50 ms | llamafile 0.9.3 |
| Q3_K - Medium | 854 | 32,354 | LocalScore ↗ | 512+128 | 70 ms | llamafile 0.9.3 |
| IQ3_S mix - 3.66 bpw | 728 | 21,230 | LocalScore ↗ | 512+128 | 69 ms | llamafile 0.9.3 |
| IQ4_XS - 4.25 bpw | 674 | 15,970 | LocalScore ↗ | 512+128 | 83 ms | llamafile 0.9.3 |
| IQ4_NL - 4.5 bpw | 651 | 25,663 | LocalScore ↗ | 512+128 | 66 ms | llamafile 0.9.3 |
| Q2_K - Medium | 646 | 21,761 | LocalScore ↗ | 512+128 | 61 ms | llamafile 0.9.3 |
| IQ4_XS - 4.25 bpw | 625 | 30,058 | LocalScore ↗ | 512+128 | 57 ms | llamafile 0.9.3 |
| IQ3_S - 3.4375 bpw | 617 | 23,654 | LocalScore ↗ | 512+128 | 57 ms | llamafile 0.9.3 |
| IQ3_XS - 3.3 bpw | 617 | 24,298 | LocalScore ↗ | 512+128 | 58 ms | llamafile 0.9.3 |
| Q3_K - Small | 557 | 14,647 | LocalScore ↗ | 512+128 | 87 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 4050 Laptop GPU 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.