GPUs & Macs › NVIDIA RTX 4050 Laptop
NVIDIA RTX 4050 Laptop: local LLM speed
Llama 2 7B (Q4_0) generates 43.7 tokens/s on the NVIDIA RTX 4050 Laptop (CUDA), and processes prompts at 1,726 tokens/s.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 2 7B | 43.7 | 1,726 | Q4_0 | CUDA FA off | tg128 | llama.cpp #15013 details |
| Llama 2 7B | 41.9 | 1,154 | Q4_0 | Vulkan FA off | tg128 | llama.cpp #10879 details |
| Llama 3.2 1B Instruct | 118 | 4,917 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
| Llama 3.1 8B Instruct | 31.4 | 1,201 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
What fits on the NVIDIA RTX 4050 Laptop
Estimates for this chip are scaled from a measured run: Llama 2 7B Q4_0 at 43.7 tok/s (CUDA, 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 | 39.7 estimate |
| Qwen3.5-9B | UD-IQ3_XXS | 4.0 GB | 41.6 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 37.3 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q3_K_M | 4.5 GB | 37.3 estimate |
| gemma-4-E4B-it | Q3_K_M | 4.1 GB | 41.2 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 39.0 estimate |
| gemma-3-1b-it | Q8_0 | 1.1 GB | 139 estimate |
| gemma-4-E2B-it | Q4_0 | 2.8 GB | 58.8 estimate |
| DeepSeek-R1-0528-Qwen3-8B | Q3_K_M | 4.1 GB | 40.5 estimate |
| Llama-3.2-3B-Instruct | Q8_0 | 3.4 GB | 48.8 estimate |
| Phi-4-mini-instruct | Q8_0 | 4.1 GB | 40.9 estimate |
| gemma-3-4b-it | Q8_0 | 4.1 GB | 40.5 estimate |
| Qwen3-14B | UD-IQ1_M | 4.1 GB | 41.1 estimate |
| gemma-3-12b-it | UD-IQ2_M | 4.4 GB | 38.3 estimate |
Similar hardware
NVIDIA RTX 3060 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