Can deepseek vl2 tiny run on the RTX 4080?
Yes. At 8K context the best fit is F16, using about 7.4 GB of the 15.2 GB usable on the RTX 4080. Estimated speed ceiling: ~305 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | Max tok/s (est., 8K) |
|---|---|---|
| 2-bit (Q2_K) | 2.1 GB Fits | 1627 |
| 3-bit (Q3_K_M) | 2.4 GB Fits | 1248 |
| 4-bit (Q4_K_M) | 2.8 GB Fits | 998 |
| 5-bit (Q5_K_M) | 3.1 GB Fits | 856 |
| 6-bit (Q6_K) | 3.5 GB Fits | 744 |
| 8-bit (Q8_0) | 4.2 GB Fits | 574 |
| 16-bit (F16/BF16) | 7.2 GB Fits | 305 |
• real GGUF file size. Speed is a bandwidth ceiling (717 GB/s), not a benchmark. Methodology.
More
deepseek vl2 tiny on every GPU and Mac · Every model that runs on the RTX 4080 · Other models that fit 16 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (deepseek-ai/deepseek-vl2-tiny), NVIDIA specs. Data as of 2026-11-17.