Can deepseek vl2 tiny run on the A100 40GB?
Yes. At 8K context the best fit is F16, using about 7.4 GB of the 38.0 GB usable on the A100 40GB. Estimated speed ceiling: ~661 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | Max tok/s (est., 8K) |
|---|---|---|
| 2-bit (Q2_K) | 2.1 GB Fits | 3528 |
| 3-bit (Q3_K_M) | 2.4 GB Fits | 2707 |
| 4-bit (Q4_K_M) | 2.8 GB Fits | 2164 |
| 5-bit (Q5_K_M) | 3.1 GB Fits | 1857 |
| 6-bit (Q6_K) | 3.5 GB Fits | 1613 |
| 8-bit (Q8_0) | 4.2 GB Fits | 1245 |
| 16-bit (F16/BF16) | 7.2 GB Fits | 661 |
• real GGUF file size. Speed is a bandwidth ceiling (1555 GB/s), not a benchmark. Methodology.
More
deepseek vl2 tiny on every GPU and Mac · Every model that runs on the A100 40GB · Other models that fit 48 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.