WeVisDoc Guide: VRAM Requirements, GGUF Builds

September 27, 2026

About This Model

WeVisDoc is a model from Tencent that turns images of document pages into structured text. Give it a PDF page or a scanned image and it returns Markdown that preserves headings and paragraph structure, with formulas as LaTeX and tables as HTML. Rather than chaining separate tools for layout detection, text recognition and table recognition, it does the whole job end-to-end in a single model.

It is built by adapting Qwen3-VL (2B and 4B), a language model that can read images, for document parsing, and comes in two sizes: WeVisDoc-2B and WeVisDoc-4B.

What Makes It Stand Out

The publisher reports that it ranks first among the compared end-to-end document parsers in all four evaluation settings. An excerpt:

Model Size OmniDocBench v1.6 overall ↑ PureDocBench real-world degraded ↑
WeVisDoc-4B 4B 95.38 69.08
WeVisDoc-2B 2B 95.06 65.60
Logics-Parsing-v2 4B 93.33 67.64
FD-RL 4B 91.21 67.04
HunyuanOCR-1.5 1B 94.74 61.59
Qianfan-OCR 4B 93.90 45.06
dots.ocr 3B 90.77 55.68
DeepSeek-OCR 2 3B 90.25 43.60
olmOCR-2-7B 7B 85.51 56.10
  • On clean documents, the top models are now close. The leaders cluster at 94–95 overall on OmniDocBench, and WeVisDoc’s lead is under one point.
  • Models separate on photographed and degraded documents. On PureDocBench’s “Real Degraded" track, Qianfan-OCR—close on clean documents—drops to 45.06 while WeVisDoc-4B holds 69.08. Some models do come close here, though: Logics-Parsing-v2 (67.64) and FD-RL (67.04) are within about 1.4–2 points. The accurate reading is “the best among the models that handle degraded documents well."
  • On clean documents, the 2B version nearly matches the 4B. Its overall 95.06 is close to the 4B’s 95.38. On degraded documents, however, it scores 65.60, behind the 4B and the two models above. If you mostly process phone photos or old scans, the 4B version is the safer choice.

Some baseline scores were re-measured by the publisher; the rest are taken from each model’s paper.

Running It Locally

  • At 4B and 2B, it is small enough for common GPUs. The official instructions run it as a vLLM server that you send page images to. This site also tracks converted builds such as GGUF (see “Quantized and Converted Variants" below).
  • It is specialized for reading documents; it is not a general-purpose chat model.
  • The license is Apache-2.0, which allows commercial use. The technical report is arXiv:2609.20423.

Source: tencent/WeVisDoc-4B model card, as of 2026-09-25. Benchmark figures are as published by Tencent.

Our Coverage and Data

Everything Local Model Watch has published about the WeVisDoc family: 1 article(s) covering the base model and its fine-tunes, plus converted builds we tracked after publication. Memory requirements below are computed by this site from file sizes, not quoted from model cards. Part of our model family index.

At a Glance

Item Value
Base model(s) tencent/WeVisDoc-4B
Publisher Tencent
Parameters 4.4B
License (model card) apache-2.0
Smallest VRAM tier 4GB
Articles 1

Hardware Requirements

Estimated requirements (calculated by Local Model Watch) — 4.4B parameters

Your VRAM Quantization File size Est. memory needed
4GB (laptop iGPU / phone class) Q5_K_S 2.9GB 3.5GB
8GB (RTX 4060 / 3060 Ti, etc.) Q8_0 4.4GB 5.2GB
12GB (RTX 4070 / 3060 12GB, etc.) F16 8.2GB 9.9GB

Memory estimates add a 20% runtime overhead (KV cache, etc.) to the actual size of the distributed files. Actual usage varies with context length, batch size and inference engine. These figures are computed by this site from file sizes, not published by the model’s authors. File sizes are measured from the converted build mradermacher/WeVisDoc-4B-GGUF. Compare with other models in our VRAM quick reference. What the quantization names mean: glossary.

Can You Run It Locally?

The publisher distributes this model as safetensors.

License — apache-2.0 (Commercial use allowed): Permits commercial use, modification and redistribution. Redistribution requires including the license and stating changes; includes a patent grant.

Compression: the Q5_K_S build measures 5.61 bits per weight — about 35% the size of the original 16-bit weights, calculated by this site from the actual file sizes.

Compiled by this site’s code from the published formats and the license field. License summaries are not legal advice — check the publisher’s original terms before relying on them.

Quantized and Converted Variants

→ Scroll horizontally to see all columns

Added Publisher Format Repository Smallest VRAM tier (build, est. memory)
2026-09-20 mradermacher GGUF mradermacher/WeVisDoc-4B-GGUF Q5_K_S 3.5GB (fits in 4GB VRAM)
2026-09-20 mradermacher GGUF (imatrix) mradermacher/WeVisDoc-4B-i1-GGUF Q5_K_S 3.5GB (fits in 4GB VRAM)

File sizes of each build:

  • Available builds in mradermacher/WeVisDoc-4B-GGUF: Q2_K 1.7GB / Q3_K_S 1.9GB / Q3_K_M 2.1GB / Q3_K_L 2.2GB / IQ4_XS 2.3GB / Q4_K_S 2.4GB / Q4_K_M 2.5GB / Q5_K_S 2.9GB / Q5_K_M 2.9GB / Q6_K 3.4GB / Q8_0 4.4GB / F16 8.2GB
  • Available builds in mradermacher/WeVisDoc-4B-i1-GGUF: IQ1_S 1.1GB / IQ1_M 1.2GB / IQ2_XXS 1.3GB / IQ2_XS 1.4GB / IQ2_S 1.5GB / IQ2_M 1.6GB / Q2_K_S 1.6GB / Q2_K 1.7GB / IQ3_XXS 1.7GB / IQ3_XS 1.8GB / Q3_K_S 1.9GB / IQ3_S 1.9GB / IQ3_M 2.0GB / Q3_K_M 2.1GB / Q3_K_L 2.2GB / IQ4_XS 2.3GB / Q4_0 2.4GB / IQ4_NL 2.4GB / Q4_K_S 2.4GB / Q4_K_M 2.5GB / Q4_1 2.6GB / Q5_K_S 2.9GB / Q5_K_M 2.9GB / Q6_K 3.4GB

This section is appended automatically by Local Model Watch when a converted build of this model appears after publication. Memory figures are estimated from the size of the distributed files.

Articles (the family’s own models first, then newest)

Published Model Type Article
2026-09-18 tencent/WeVisDoc-4B New Models Tencent Releases WeVisDoc-2B and WeVisDoc-4B Document Parsing Models

Repositories

Last updated 2026-09-27 (JST). The explanation at the top of this page was written with the help of AI from the primary sources it cites. The tables and lists under “Our Coverage and Data" are assembled by code from our article log.