GPUs & Macs › NVIDIA RTX 4060
NVIDIA RTX 4060: local LLM speed
6 measured results for the NVIDIA RTX 4060.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 3.2 1B Instruct | 134 | 5,352 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
| gpt-oss 20B | 42.3 | 1,642 | MXFP4 | CUDA FA on | tg128 · part CPU | llama.cpp #15396 details |
| Llama 3.1 8B Instruct | 38.3 | 1,544 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
| gpt-oss 20B | 32.7 | 1,430 | F16 | CUDA FA on | tg128 · part CPU | llama.cpp #15396 details |
| gpt-oss 120B | 18.9 | 333 | MXFP4 | CUDA FA on | tg128 · part CPU | llama.cpp #15396 details |
| gpt-oss 120B | 17.4 | 201 | F16 | CUDA FA on | tg128 · part CPU | llama.cpp #15396 details |
What fits on the NVIDIA RTX 4060
Estimates for this chip are scaled from a measured run: Llama 3.1 8B Instruct Q4_K_M at 38.3 tok/s (llamafile, 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 | IQ4_XS | 6.4 GB | 29.6 estimate |
| Qwen3.5-9B | Q5_K_S | 6.4 GB | 29.6 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 42.1 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q4_K_M | 5.8 GB | 32.3 estimate |
| gemma-4-E4B-it | Q5_K_M | 5.5 GB | 34.4 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 44.0 estimate |
| gemma-3-1b-it | Q8_0 | 1.1 GB | 157 estimate |
| gemma-4-E2B-it | Q8_0 | 5.0 GB | 37.9 estimate |
| DeepSeek-R1-0528-Qwen3-8B | Q5_K_M | 5.9 GB | 32.2 estimate |
| Llama-3.2-3B-Instruct | Q8_0 | 3.4 GB | 55.1 estimate |
| Phi-4-mini-instruct | Q8_0 | 4.1 GB | 46.1 estimate |
| gemma-3-4b-it | Q8_0 | 4.1 GB | 45.6 estimate |
| Qwen3-14B | UD-IQ3_XXS | 6.0 GB | 31.3 estimate |
| gemma-3-12b-it | Q3_K_M | 6.0 GB | 31.4 estimate |
Similar hardware
NVIDIA RTX 2070 SUPER · NVIDIA RTX 4060 Ti · NVIDIA RTX 2060 SUPER · NVIDIA GTX 1070 Ti · NVIDIA RTX 3060 Ti · NVIDIA RTX 3070 · NVIDIA GTX 1070 · NVIDIA RTX 3070 Ti