RTX 3060 Ti × Llama 3.2 1B Instruct
Llama 3.2 1B Instruct on RTX 3060 Ti
244best gen tok/s
8,269best prompt tok/s
640max context tested
9results
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 | 244 | 7,658 | LocalScore ↗ | 512+128 | 181 ms | llamafile 0.9.2 |
| Q4_K - Medium | 242 | 7,283 | LocalScore ↗ | 512+128 | 194 ms | llamafile 0.9.2 |
| Q4_K - Medium | 199 | 7,575 | LocalScore ↗ | 512+128 | 180 ms | llamafile 0.9.3 |
| Q4_K - Medium | 195 | 7,724 | LocalScore ↗ | 512+128 | 178 ms | llamafile 0.9.3 |
| Q4_K - Medium | 182 | 8,050 | LocalScore ↗ | 512+128 | 169 ms | llamafile 0.9.2 |
| Q4_K - Medium | 178 | 8,269 | LocalScore ↗ | 512+128 | 168 ms | llamafile 0.9.3 |
| Q4_K - Medium | 173 | 7,408 | LocalScore ↗ | 512+128 | 185 ms | llamafile 0.9.2 |
| Q4_K - Medium | 169 | 7,163 | LocalScore ↗ | 512+128 | 191 ms | llamafile 0.9.2 |
| Q4_K - Medium | 153 | 7,662 | LocalScore ↗ | 512+128 | 180 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 3060 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.