What GPU do I need to run deepseek-ai/DeepSeek-V3.2-Exp?
A 685B (MoE) language model for chat and instruction-following. 685.4B parameters, published in F8_E4M3. View on Hugging Face
DeepSeek-V3.2-Exp is published by deepseek-ai on Hugging Face, with 82,638 downloads and 999 likes to date. It's a DeepseekV32ForCausalLM model built for text-generation, published natively in F8_E4M3.
What DeepSeek-V3.2-Exp is
DeepSeek-V3.2-Exp is a 685B-parameter mixture-of-experts language model published by DeepSeek on Hugging Face. It is released under MIT.
License note: permissive: allows commercial use, modification and redistribution. Facts in this section are sourced from DeepSeek-V3.2-Exp's Hugging Face model card, not benchmarked by Aquanode.
What it's used for
- Chat assistants
- Instruction following
- Synthetic data generation
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision, plus a fixed overhead for KV-cache, activations, and fragmentation. Full formula and assumptions: methodology.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| FP8 (native) | 638.3 GB | 766.0 GB | RTX PRO 6000 WS | 8 | $11.98/hr |
| INT4 (quantized) | 319.2 GB | 383.0 GB | RTX A6000 | 8 | $2.90/hr |
A GPU is only matched to a row if its hardware supports that precision, and the primary recommendation is always a single-GPU fit when one exists.
INT4 requires a quantized checkpoint actually published for this model, check its Hugging Face page before relying on this row.
Cheapest way to run DeepSeek-V3.2-Exp at its published (F8_E4M3) precision: 8× RTX PRO 6000 WS, at $1.50/hr per GPU ($11.98/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
DeepSeek-V3.2-Exp: common questions
Can DeepSeek-V3.2-Exp run on a single GPU?
No. At FP8 (native) it needs 766.0 GB of VRAM, more than a single 32 GB desktop card holds. The cheapest capable card in the live feed is a 96.0 GB RTX PRO 6000 WS, and it takes 8 of them.
Is DeepSeek-V3.2-Exp already quantized?
Yes. It is published in FP8, one byte per parameter, so the 766.0 GB figure is already a quantized footprint rather than a full-precision one. Only INT4 goes below it, at 383.0 GB. A GPU without FP8 tensor cores cannot run it as published, which is why cards here are matched on precision support and not on VRAM alone.
How many GPUs do I need to run DeepSeek-V3.2-Exp?
8 at FP8 (native). It needs 766.0 GB of VRAM and the cheapest capable live offer is a 96.0 GB RTX PRO 6000 WS, so 8 of them come to $11.98/hr in total.
Does quantizing DeepSeek-V3.2-Exp lower the GPU bill?
Yes. At FP8 (native) the cheapest live fit is 8 RTX PRO 6000 WS cards at $11.98/hr. At INT4 (quantized) it drops to 8 RTX A6000 cards at $2.90/hr, provided a quantized checkpoint exists for it.
How to run DeepSeek-V3.2-Exp
Run DeepSeek-V3.2-Exp with SGLang
From deepseek-ai/DeepSeek-V3.2-Exp's own deployment docs.
# H200
docker pull lmsysorg/sglang:dsv32
# MI350
docker pull lmsysorg/sglang:dsv32-rocm
# NPUs
docker pull lmsysorg/sglang:dsv32-a2
docker pull lmsysorg/sglang:dsv32-a3Source: https://huggingface.co/deepseek-ai/DeepSeek-V3.2-Exp/raw/main/README.md
Deploy DeepSeek-V3.2-Exp on Aquanode
Aquanode has no one-click deploy template for DeepSeek-V3.2-Exp; you install the inference engine yourself with the commands below. Aquanode sells GPU pods billed per second, not a hosted inference API.
- Launch a bare GPU pod sized to the requirement above (8× RTX PRO 6000 WS or larger).
- Open a terminal on the pod, or save one of the commands above as a startup script so it runs automatically the first time the pod boots.
- Run the command and connect to the resulting endpoint.
Launch a GPU podWeight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
More DeepSeek V3 models
- DeepSeek-V3.2 (685.4B, F8_E4M3)
- DeepSeek-V3-0324 (684.5B, F8_E4M3)
- DeepSeek-V3 (684.5B, F8_E4M3)
- DeepSeek-V3.1 (684.5B, F8_E4M3)
- DeepSeek-V3.1-Base (684.5B, F8_E4M3)
Related reading: RTX PRO 6000 WS pricing and specs, The NVIDIA Inception program, explained, Google Colab alternatives for dedicated GPU access, Free GPU credits for students and researchers, and RunPod volume disk vs. network volume, compared.
