MinerU VRAM requirements
OpenDataLab's open document parser, which turns PDFs, scans and Office files into Markdown, HTML or JSON. Its Basic tier runs small OCR and layout models; its Standard and Advanced tiers add a vision language model for harder layouts. OpenDataLab's figures run from at least 4 GB for MinerU 4.0 Basic to at least 8 GB for MinerU 4.0 Standard or Advanced. The cheapest single GPU with 8 GB or more holds all of them: the 16 GB Tesla V100 at Verda (DataCrunch), $0.26 an hour.
What OpenDataLab says each setting needs
| Model and setting | Stated GPU memory | Cheapest single GPU with that much | Per hour | |
|---|---|---|---|---|
| MinerU 4.0 Standard or AdvancedPyTorch with vLLM or LMDeploy (the full extra), on a Volta-or-newer NVIDIA GPU | at least 8 GB | Tesla V100, 16 GB, Verda (DataCrunch) | $0.26 | Rent → |
| MinerU 4.0 BasicPyTorch on an NVIDIA GPU (the torch extra) | at least 4 GB | Tesla V100, 16 GB, Verda (DataCrunch) | $0.26 | Rent → |
The memory figures are OpenDataLab's, not our estimate. The price is the lowest on-demand rate for one GPU with at least that much memory, from our latest check; a card that holds a model is not a promise of its speed.
Before you rent
The README asks for that much available GPU memory, plus at least 16 GB of system memory for the full extra. The Basic tier also runs on a CPU with no GPU at all, so a rented GPU buys throughput on a large batch of documents rather than making MinerU work.
In OpenDataLab's words
- MinerU 4.0 Standard or Advanced: “A Volta-or-newer NVIDIA GPU with at least 8 GB available VRAM.”
- MinerU 4.0 Basic: “| Linux/Windows with NVIDIA GPU and 4GB+ VRAM | `basic` | PyTorch | `torch` |”
Source: the MinerU README on GitHub, read 17 October 2026.
Text models are sized by a formula instead: LLM VRAM requirements for every open language model, or the LLM VRAM calculator for your own context and batch.
More models sized by their publishers' figures
- Whisper (speech recognition, OpenAI)
- Stable Diffusion 3.5 Medium (image generation, Stability AI)
- Wan2.1 (video generation, Wan-AI)
- Wan2.2 (video generation, Wan-AI)
- HunyuanVideo (video generation, Tencent)
- olmOCR (OCR and PDF conversion, Ai2)
- GTE-Qwen2 (text embeddings, Alibaba)