▸ DEVICE UNDER TEST
NVIDIA P102-100 — 5 GB VRAM.
▸ P102-100 SPEC
- BRAND
- NVIDIA
- VRAM
- 5 GB GDDR5X
- BANDWIDTH
- 440 GB/s
- FP16 COMPUTE
- 0.2 TFLOPS
- FP32 COMPUTE
- 10.8 TFLOPS
- CUDA CORES
- 3,200
- TDP
- 250 W
- ARCHITECTURE
- Pascal
▸ AI CAPABILITY
80/ 331 models @ Q4
With 5 GB VRAM and 440 GB/s bandwidth, this GPU handles models up to 6.24B parameters.
Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~50 tok/s.
§ 01TOP MODELS FOR P102-100
80 FIT · SHOWING 20| MODEL | SIZE | VRAM Q4 | TOK/S | AVG |
|---|---|---|---|---|
| ChatGLM2 6B | 6.24B | 4.3 GB | 56 | 20.7 |
| ChatGLM3 6B | 6.24B | 4.3 GB | 56 | 42.7 |
| Yi-1.5 6B | 6.06B | 4.2 GB | 58 | 28.6 |
| Gemma 3n E2B | 6B | 4.2 GB | 59 | 21.7 |
| Yi 6B | 6B | 4.2 GB | 59 | 23.2 |
| Gemma 4 E2B | 5.1B | 3.6 GB | 69 | 25.7 |
| Qwen3.5-4B | 4.7B | 3.4 GB | 75 | 29.3 |
| InternLM2 5B | 4.5B | 3.2 GB | 78 | 47.6 |
| Gemma 3 4B | 4.3B | 3.1 GB | 82 | 22.8 |
| TranslateGemma 4B | 4B | 2.9 GB | 88 | 5.5 |
| MedGemma 1.5 4B | 4B | 2.9 GB | 88 | 5.5 |
| Qwen 1.5 4B | 4B | 2.9 GB | 88 | 12.6 |
| Qwen3 4B | 4B | 2.9 GB | 88 | 40.7 |
| Qwen 3.5 4B | 4B | 2.9 GB | 88 | 47.4 |
| Nemotron 3 Nano 4B | 3.97B | 2.9 GB | 89 | 32.0 |
| Phi-3.5 Mini 3.8B | 3.82B | 2.8 GB | 92 | 46.6 |
| phi-3-mini-4k 3.8B | 3.8B | 2.8 GB | 93 | 30.5 |
| Phi-4-mini 3.8B | 3.8B | 2.8 GB | 93 | 49.0 |
| Qwen2.5-VL-3B | 3.8B | 2.8 GB | 93 | 29.9 |
| granite-4.0-h-micro 3.2B | 3.2B | 2.4 GB | 110 | 18.4 |