GPUs & Macs › NVIDIA L4
NVIDIA L4: local LLM speed
3 measured results for the NVIDIA L4.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 3.2 1B Instruct | 188 | 9,268 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
| Llama 3.1 8B Instruct | 42.9 | 2,056 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
| Qwen2.5 14B Instruct | 23.6 | 1,095 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
What fits on the NVIDIA L4
Estimates for this chip are scaled from a measured run: Qwen2.5 14B Instruct Q4_K_M at 23.6 tok/s (llamafile, source). Fit assumes the whole model in GPU memory with ~8K context.
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| Qwen3.8-27B | UD-Q6_K | 22.0 GB | 9.6 estimate |
| gemma-4-12B-it | Q8_0 | 12.7 GB | 16.7 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 22.3 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 47.3 estimate |
| Qwen3.6-35B-A3B | UD-Q4_K_S | 20.9 GB | 59.2 estimate |
| gemma-4-26B-A4B-it | UD-Q5_K_M | 21.2 GB | 32.6 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 22.2 estimate |
| gemma-4-E4B-it | Q8_0 | 8.2 GB | 25.9 estimate |
| Muse-Glimmer-30B | UD-Q5_K_M | 19.2 GB | 11.0 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 49.5 estimate |
| gemma-4-31B-it | Q5_K_M | 21.7 GB | 9.8 estimate |
| gemma-4-26B-A4B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 14.2 GB | 48.4 estimate |
| gemma-4-12B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 6.7 GB | 31.6 estimate |
| gemma-3-1b-it | Q8_0 | 1.1 GB | 177 estimate |
Similar hardware
NVIDIA A30 · NVIDIA Tesla P40 · NVIDIA Tesla M40 · NVIDIA A10 · NVIDIA Tesla V100 32GB · NVIDIA Tesla P100 · NVIDIA Tesla T4 · NVIDIA Tesla V100 16GB