GPUs & Macs › NVIDIA Quadro P1000
NVIDIA Quadro P1000: local LLM speed
Llama 2 7B (Q4_0) generates 14.0 tokens/s on the NVIDIA Quadro P1000 (CUDA), and processes prompts at 183 tokens/s.
Measured results
| Model | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Llama 2 7B | 14.0 | 183 | Q4_0 | CUDA FA off | tg128 | llama.cpp #15013 details |
| Llama 2 7B | 13.7 | 174 | Q4_0 | CUDA FA on | tg128 | llama.cpp #15013 details |
What fits on the NVIDIA Quadro P1000
Estimates for this chip are scaled from a measured run: Llama 2 7B Q4_0 at 14.0 tok/s (CUDA, source). Fit assumes the whole model in GPU memory with ~8K context.
| Model | Best quant that fits | File | Expected tok/s |
|---|---|---|---|
| Qwen3.5-4B | IQ4_XS | 2.5 GB | 21.6 estimate |
| Qwen3-4B | Q4_K_M | 2.5 GB | 21.4 estimate |
| gemma-3-1b-it | Q8_0 | 1.1 GB | 44.6 estimate |
| DeepSeek-R1-0528-Qwen3-8B | UD-IQ1_M | 2.4 GB | 22.4 estimate |
| Llama-3.2-3B-Instruct | Q5_K_M | 2.3 GB | 23.0 estimate |
| Phi-4-mini-instruct | Q4_K_M | 2.5 GB | 21.5 estimate |
| gemma-3-4b-it | Q4_K_M | 2.5 GB | 21.5 estimate |
| Qwen3.5-0.8B | Q8_0 | 0.8 GB | 44.6 estimate |
| Llama-3.2-3B-Instruct-uncensored | Q5_K_M | 2.6 GB | 20.6 estimate |
| Qwen3-0.6B | Q8_0 | 0.8 GB | 44.6 estimate |
| GLM-OCR | Q8_0 | 1.0 GB | 44.6 estimate |
| Qwen3-ASR-1.7B | Q8_0 | 2.2 GB | 24.7 estimate |
| SmolLM3-3B | Q4_K_M | 1.9 GB | 27.9 estimate |
| Nanbeige4.2-3B | Q4_K_S | 2.6 GB | 20.9 estimate |
Similar hardware
NVIDIA T1000 · NVIDIA Quadro P2000 · NVIDIA RTX A2000 · NVIDIA Quadro RTX 4000 · NVIDIA RTX A2000 12GB · NVIDIA RTX A4000 · NVIDIA RTX 2000 Ada · NVIDIA RTX A4500