Reasoning

What GPU do I need to run deepseek-ai/DeepSeek-V4-Pro-0813?

A 1.6T-parameter (49B active) reasoning model with a 1M-token context. 1650.5B parameters, published in FP4 + FP8 Mixed. View on Hugging Face

1650.5B
Parameters
FP4 + FP8 Mixed
Native precision
1M tokens (1,048,576)
Context length
MIT
License
Text
Modality
DeepSeek
Organization
Mixture-of-experts: 49B active parameters (stated on the model card)
Active parameters (MoE)

What DeepSeek-V4-Pro-0813 is

DeepSeek-V4-Pro-0813 is DeepSeek's official release of DeepSeek-V4-Pro, a 1.6T-parameter mixture-of-experts reasoning model with 49B active parameters and a DSpark speculative-decoding module attached, superseding the preview version with improved agentic performance. Its own model card states a 1,048,576-token (1M) context window via a hybrid Compressed/Heavily-Compressed attention architecture, three reasoning-effort levels (low, high, max), and mixed FP4 (MoE experts) + FP8 (other parameters) published precision.

License note: fully permissive, no gating and no usage restrictions. Facts in this section are sourced from DeepSeek-V4-Pro-0813's Hugging Face model card, not benchmarked by Aquanode.

What it's used for

  • Multi-step reasoning and agentic coding
  • Long-context analysis (up to 1M tokens)
  • Tool use and terminal/agent tasks

Benchmarks (published by DeepSeek)

Published by DeepSeek, not measured by Aquanode.

VRAM required & cheapest live GPU fit

Required VRAM = weight size at each precision times a flat 1.2 overhead for activations and fragmentation. The KV-cache grows with context and is not in that factor; it is listed per model below where the architecture is published. Full formula and assumptions: methodology.

PrecisionWeight sizeRequired VRAMCheapest live fitGPUs neededEst. $/hr (full fit)
FP4 + FP8 Mixed (native)831.4 GB1128.0 GBAMD MI300X6$15.77/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.

The SGLang DeepSeek-V4 cookbook linked from DeepSeek's model card lists DeepSeek-V4-Pro-0813 as verified on 8×H200 (FP4, tensor-parallel-8); NVIDIA's own H200 spec lists 141GB HBM3e per GPU, so 8×141GB ≈ 1128GB total.

Cheapest way to run DeepSeek-V4-Pro-0813 at its published (FP4 + FP8 Mixed) precision: 6× AMD MI300X, at $2.63/hr per GPU ($15.77/hr total).

DeepSeek-V4-Pro-0813: KV cache by context length

The KV cache is the memory the attention layers hold for every token of context, on top of the weights. It is not part of the flat 1.2x overhead in the table above, grows with context length and with every concurrent request, and is why a long-context deployment needs more VRAM than the table shows.

KV-cache size for DeepSeek-V4-Pro-0813 is not published for this architecture, so no figure is shown.

How to run DeepSeek-V4-Pro-0813

Run DeepSeek-V4-Pro-0813 with vLLM

From DeepSeek-V4-Pro-0813's own model card: an example serving it with vLLM and DSpark speculative decoding on a single 4×GB300 node.

vllm serve deepseek-ai/DeepSeek-V4-Pro-0813 \
  --trust-remote-code --kv-cache-dtype fp8 --block-size 256 \
  --data-parallel-size 4 --enable-expert-parallel \
  --moe-backend deep_gemm_mega_moe \
  --attention-config '{"use_fp4_indexer_cache": true}' \
  --speculative-config '{"method":"dspark","num_speculative_tokens":7,"draft_sample_method":"greedy"}'

Source: https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813/raw/main/README.md

Run DeepSeek-V4-Pro-0813 with SGLang

From DeepSeek-V4-Pro-0813's own model card, same single 4×GB300 node example as the vLLM command above.

sglang serve \
  --trust-remote-code \
  --model-path deepseek-ai/DeepSeek-V4-Pro-0813 \
  --tp 4 \
  --moe-runner-backend flashinfer_mxfp4 \
  --speculative-algorithm DSPARK \
  --mem-fraction-static 0.90 \
  --chunked-prefill-size 4096

Source: https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813/raw/main/README.md

Deploy DeepSeek-V4-Pro-0813 on Aquanode

Aquanode has no one-click deploy template for DeepSeek-V4-Pro-0813; you install the inference engine yourself with the commands below. Aquanode sells GPU pods billed per second, not a hosted inference API.

  1. Launch a bare GPU pod sized to the requirement above (6× AMD MI300X or larger).
  2. 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.
  3. Run the command and connect to the resulting endpoint.
Launch a GPU pod

Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.

More DeepSeek V4 models

All 3 DeepSeek V4 models: VRAM and GPU requirements

Alternatives at this size

Other models for text-generation within about a third of DeepSeek-V4-Pro-0813's 1650.5B parameters, from other model lines.

More on DeepSeek-V4-Pro-0813

Related reading: AMD MI300X pricing and specs, How much VRAM you need for LLMs, Serving LLMs with vLLM, vLLM vs TensorRT-LLM vs SGLang, and Best GPU for LLM inference.

Your own GPU cloud accounts

One console for the GPU clouds you already rent from. Your keys stay on your machine.

$20 a month, flat

Connect your accounts

One console for the GPU clouds you already use.

Connect your accounts for $20 a month, flat. Your provider bills the GPU.

© 2026 Aquanode. All rights reserved.

All trademarks, logos and brand names are the property of their respective owners.