▸ DEVICE UNDER TEST
NVIDIA GeForce GT 640 OEM Rebrand — 2 GB VRAM.
▸ GEFORCE GT 640 OEM REBRAND SPEC
- BRAND
- NVIDIA
- VRAM
- 2 GB DDR3
- BANDWIDTH
- 38.4 GB/s
- FP16 COMPUTE
- 0.4 TFLOPS
- FP32 COMPUTE
- 0.4 TFLOPS
- CUDA CORES
- 144
- TDP
- 75 W
- ARCHITECTURE
- Fermi 2.0
▸ AI CAPABILITY
37/ 449 models @ Q4
With 2 GB VRAM and 38.4 GB/s bandwidth, this GPU handles models up to 1.35B parameters.
Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~4 tok/s.
§ 01TOP MODELS FOR GEFORCE GT 640 OEM REBRAND
37 FIT · SHOWING 20| MODEL | SIZE | VRAM Q4 | TOK/S | AVG |
|---|---|---|---|---|
| DeepSeek Coder 1.3B | 1.35B | 1.3 GB | 23 | 16.8 |
| EXAONE-4.0-1.2B | 1.3B | 1.3 GB | 24 | 18.9 |
| OPT 1.3B | 1.3B | 1.3 GB | 24 | 5.3 |
| MiniCPM-V 4.6 | 1.3B | 1.3 GB | 24 | 21.5 |
| LFM2.5-1.2B-Thinking | 1.2B | 1.2 GB | 26 | 19.6 |
| Llama-3.2-1B | 1.2B | 1.2 GB | 26 | 10.1 |
| LFM2 1.2B | 1.2B | 1.2 GB | 26 | 15.6 |
| Zamba2 1.2B | 1.2B | 1.2 GB | 26 | 41.5 |
| TinyLlama 1.1B | 1.1B | 1.2 GB | 28 | 13.6 |
| MiniCPM5 1B | 1.08B | 1.1 GB | 28 | 26.5 |
| InternVL3 1B | 0.94B | 1.1 GB | 33 | 58.0 |
| Qwen 3.5 0.8B | 0.87B | 1.0 GB | 35 | 22.5 |
| GPT-2 Large 774M | 0.81B | 1.0 GB | 38 | 5.6 |
| Qwen3 0.6B | 0.75B | 0.9 GB | 41 | 19.1 |
| LFM2 700M | 0.74B | 0.9 GB | 42 | 50.4 |
| Qwen 1.5 0.5B | 0.62B | 0.9 GB | 50 | 9.9 |
| Qwen3-Embedding 0.6B | 0.6B | 0.9 GB | 51 | — |
| Falcon-H1R Tiny 0.6B | 0.6B | 0.9 GB | 51 | — |
| Falcon Perception 0.6B | 0.6B | 0.9 GB | 51 | — |
| BGE-M3 | 0.568B | 0.8 GB | 54 | 63.0 |