Can WeVisOne VL 2B run on the L40S?
Yes. At 8K context the best fit is F16, using about 5.4 GB of the 45.6 GB usable on the L40S. Estimated speed ceiling: ~212 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | 128K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|---|
| 2-bit (Q2_K) • | 1.9 GB Fits | 2.4 GB Fits | 5.0 GB Fits | 15.5 GB Fits | 980 |
| 3-bit (Q3_K_M) • | 2.1 GB Fits | 2.6 GB Fits | 5.2 GB Fits | 15.7 GB Fits | 804 |
| 4-bit (Q4_K_M) • | 2.3 GB Fits | 2.8 GB Fits | 5.4 GB Fits | 15.9 GB Fits | 673 |
| 5-bit (Q5_K_M) • | 2.5 GB Fits | 2.9 GB Fits | 5.6 GB Fits | 16.1 GB Fits | 586 |
| 6-bit (Q6_K) • | 2.7 GB Fits | 3.1 GB Fits | 5.8 GB Fits | 16.3 GB Fits | 516 |
| 8-bit (Q8_0) • | 3.1 GB Fits | 3.6 GB Fits | 6.2 GB Fits | 16.7 GB Fits | 398 |
| 16-bit (F16/BF16) • | 4.9 GB Fits | 5.4 GB Fits | 8.0 GB Fits | 18.5 GB Fits | 212 |
• real GGUF file size from mradermacher/WeVisOne-VL-2B-GGUF. Speed is a bandwidth ceiling (864 GB/s), not a benchmark. Methodology.
More
WeVisOne VL 2B on every GPU and Mac · Every model that runs on the L40S · Other models that fit 48 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (tencent/WeVisOne-VL-2B), NVIDIA specs. Data as of 2026-11-17.