Quadro P2000 Mobile holds 96 of the models in our catalogue and is, in practice, a Q4_K_M card — the largest it takes is Ministral 3 3B Reasoning at Q4_K_M.
Neighbours in memory rather than in price: memory decides whether a card can do the job at all, so two cards of the same size at different prices is the comparison you are making. Ordered by memory, then bandwidth — not by the value column, which is worked out from bandwidth and price alone and therefore rewards a cheap card whatever its software stack does to that bandwidth in practice. Read it as one input, not as a ranking.
NVIDIA Quadro P2000 Mobile — 4 GB VRAM.
- BRAND
- NVIDIA
- VRAM
- 4 GB GDDR5
- BANDWIDTH
- 96.1 GB/s
- FP16 COMPUTE
- 6 TFLOPS
- FP32 COMPUTE
- 3 TFLOPS
- CUDA CORES
- 1,152
- TDP
- 75 W
- ARCHITECTURE
- Pascal
With 4 GB VRAM and 96.1 GB/s bandwidth, this GPU handles models up to 4B parameters.
Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~11 tok/s.
| MODEL | SIZE | VRAM Q4 | TOK/S | AVG |
|---|---|---|---|---|
| Qwen 1.5 4B | 4B | 2.9 GB | 19 | 12.6 |
| Qwen3 4B | 4B | 2.9 GB | 19 | 40.7 |
| Qwen3-4B Instruct 2507 | 4B | 2.9 GB | 19 | 37.2 |
| Qwen3-Embedding 4B | 4B | 2.9 GB | 19 | — |
| Nemotron 3 Nano 4B | 3.97B | 2.9 GB | 19 | 32.0 |
| Ministral 3 3B | 3.85B | 2.8 GB | 20 | 21.4 |
| Phi-3.5 Mini 3.8B | 3.82B | 2.8 GB | 20 | 46.6 |
| phi-3-mini-4k 3.8B | 3.8B | 2.8 GB | 20 | 30.5 |
| Phi-4-mini 3.8B | 3.8B | 2.8 GB | 20 | 49.0 |
| Cogito 3B | 3.61B | 2.7 GB | 21 | 22.1 |
| Falcon3-3B | 3.23B | 2.5 GB | 24 | 25.7 |
| granite-4.0-h-micro 3.2B | 3.2B | 2.4 GB | 24 | 18.4 |
| Llama-3.2-3B | 3.2B | 2.4 GB | 24 | 17.9 |
| Falcon-H1 3B | 3.15B | 2.4 GB | 24 | 49.5 |
| Qwen 2.5 3B | 3.1B | 2.4 GB | 25 | 37.2 |
| SmolLM3-3B | 3.1B | 2.4 GB | 25 | 30.5 |
| Ministral 3B | 3B | 2.3 GB | 26 | 29.6 |
| StarCoder2 3B | 3B | 2.3 GB | 26 | 9.5 |
| Granite 4.1 3B | 3B | 2.3 GB | 26 | 16.6 |
| Granite 4.0 3B Vision | 3B | 2.3 GB | 26 | 23.9 |