GPUs & Macs › NVIDIA A2
NVIDIA A2: local LLM speed
1 measured result for the NVIDIA A2.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Qwen2.5 14B Instruct | 5.6 | 160 | Q4_K_M | llamafile | LocalScore avg | LocalScore details |
What fits on the NVIDIA A2
Estimates for this chip are scaled from a measured run: Qwen2.5 14B Instruct Q4_K_M at 5.6 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-IQ3_XXS | 10.9 GB | 4.6 estimate |
| gemma-4-12B-it | Q8_0 | 12.7 GB | 3.9 estimate |
| Qwen3.5-9B | Q8_0 | 9.5 GB | 5.2 estimate |
| Qwen3.5-4B | Q8_0 | 4.5 GB | 11.2 estimate |
| Qwen3.6-35B-A3B | UD-IQ2_M | 11.5 GB | 25.3 estimate |
| gemma-4-26B-A4B-it | UD-Q3_K_M | 12.7 GB | 12.8 estimate |
| MiMo-V2.6-Distill-Qwen-9B | Q8_0 | 9.5 GB | 5.2 estimate |
| gemma-4-E4B-it | Q8_0 | 8.2 GB | 6.1 estimate |
| Muse-Glimmer-30B | UD-IQ2_M | 12.3 GB | 4.1 estimate |
| Qwen3-4B | Q8_0 | 4.3 GB | 11.7 estimate |
| gemma-4-31B-it | UD-IQ3_XXS | 11.8 GB | 4.2 estimate |
| gemma-4-12B-it-qat-q4_0-unquantized | UD-Q4_K_XL | 6.7 GB | 7.4 estimate |
| gemma-3-1b-it | Q8_0 | 1.1 GB | 41.7 estimate |
| gemma-4-E2B-it | Q8_0 | 5.0 GB | 10.1 estimate |
Similar hardware
NVIDIA RTX 5080 · NVIDIA RTX 5070 Ti · NVIDIA RTX 4080 SUPER · NVIDIA RTX 4080 · NVIDIA RTX 4070 Ti SUPER · NVIDIA RTX 5060 Ti · NVIDIA RTX PRO 2000 Blackwell · NVIDIA Quadro P5000