GPUs & Macs › NVIDIA RTX 4000 SFF Ada
NVIDIA RTX 4000 SFF Ada: local LLM speed
3 measured results for the NVIDIA RTX 4000 SFF Ada.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 3.2 1B Instruct | 189 | 8,753 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
| Llama 3.1 8B Instruct | 44.4 | 2,004 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
| Qwen2.5 14B Instruct | 24.7 | 1,051 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
What fits on the NVIDIA RTX 4000 SFF Ada
Estimates for this chip are scaled from a measured run: Qwen2.5 14B Instruct Q4_K_M at 24.7 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-Q4_K_M | 16.5 GB | 13.5 estimate |
| gemma-4-12B-it | Q8_0 | 12.7 GB | 17.5 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 23.3 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 49.5 estimate |
| Qwen3.6-35B-A3B | UD-IQ4_XS | 17.7 GB | 73.0 estimate |
| gemma-4-26B-A4B-it | UD-Q4_K_M | 16.9 GB | 42.6 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 23.2 estimate |
| gemma-4-E4B-it | Q8_0 | 8.2 GB | 27.1 estimate |
| Muse-Glimmer-30B | UD-IQ3_M | 14.1 GB | 15.7 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 51.8 estimate |
| gemma-4-31B-it | Q4_K_S | 17.4 GB | 12.8 estimate |
| gemma-4-26B-A4B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 14.2 GB | 50.6 estimate |
| gemma-4-12B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 6.7 GB | 33.0 estimate |
| gemma-3-1b-it | Q8_0 | 1.1 GB | 185 estimate |
Similar hardware
NVIDIA RTX A4500 · NVIDIA RTX 4000 Ada · NVIDIA RTX PRO 4000 Blackwell · NVIDIA RTX A5000 · NVIDIA Quadro RTX 6000 · NVIDIA RTX A4000 · NVIDIA RTX 2000 Ada · NVIDIA RTX A2000 12GB