GPUs & Macs › NVIDIA RTX 3060 Laptop
NVIDIA RTX 3060 Laptop: local LLM speed
Llama 2 7B (Q4_0) generates 49.0 tokens/s on the NVIDIA RTX 3060 Laptop (Vulkan), and processes prompts at 1,060 tokens/s.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 2 7B | 49.0 | 1,060 | Q4_0 | Vulkan FA off | tg128 | llama.cpp #10879 details |
| Llama 3.2 1B Instruct | 151 | 5,430 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
| Llama 3.1 8B Instruct | 40.8 | 1,253 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
| Qwen2.5 14B Instruct | 1.1 | 83 | Q4_K_M | llamafile | LocalScore avg · part CPU | LocalScore details |
What fits on the NVIDIA RTX 3060 Laptop
Estimates for this chip are scaled from a measured run: Llama 2 7B Q4_0 at 49.0 tok/s (Vulkan, source). Fit assumes the whole model in GPU memory with ~8K context.
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| gemma-4-12B-it | UD-IQ2_M | 4.2 GB | 44.5 estimate |
| Qwen3.5-9B | UD-IQ3_XXS | 4.0 GB | 46.7 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 41.8 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q3_K_M | 4.5 GB | 41.8 estimate |
| gemma-4-E4B-it | Q3_K_M | 4.1 GB | 46.2 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 43.8 estimate |
| gemma-3-1b-it | Q8_0 | 1.1 GB | 156 estimate |
| gemma-4-E2B-it | Q4_0 | 2.8 GB | 66.0 estimate |
| DeepSeek-R1-0528-Qwen3-8B | Q3_K_M | 4.1 GB | 45.4 estimate |
| Llama-3.2-3B-Instruct | Q8_0 | 3.4 GB | 54.8 estimate |
| Phi-4-mini-instruct | Q8_0 | 4.1 GB | 45.9 estimate |
| gemma-3-4b-it | Q8_0 | 4.1 GB | 45.4 estimate |
| Qwen3-14B | UD-IQ1_M | 4.1 GB | 46.1 estimate |
| gemma-3-12b-it | UD-IQ2_M | 4.4 GB | 42.9 estimate |
Similar hardware
NVIDIA RTX 4050 Laptop · NVIDIA RTX 1000 Ada Laptop · NVIDIA RTX 3050 6GB Laptop · NVIDIA RTX 3070 Laptop · NVIDIA RTX 4060 Laptop · NVIDIA RTX 5070 Laptop · NVIDIA RTX 3070 Ti Laptop · NVIDIA RTX 4070 Laptop