All tasks / Models
WeVisDoc (2B and 4B)
Tencent's open-weights document-parsing VLMs in 2B and 4B sizes.
- Tencent
- Open weights
- Released 16 Sep 2026
Where it ranks
- –Documents & OCRModels · Local · no result on this task’s benchmarks yet
Facts
- Vendor
- Tencent
- Family
- WeVisDoc
- Released
- 16 Sep 2026
- Weights
- Open: download and run it yourself
- License
- Apache-2.0
- Parameters
- 4.44B
- GPU memory to run locally
- ~10 GB (Estimate for the 4B variant, no official figure: BF16 weights ≈ 4.44 B × 2 GB + ~1.5 GB overhead; the 2B variant needs about half.)
- Access
- Free (open weights)
- Last verified
- 19 Oct 2026
Price from huggingface.co, 19 Oct 2026.
Benchmark results
| Benchmark | Result | Entry on the leaderboard | As of | Source |
|---|---|---|---|---|
| ParseBench (models) | 38.9% | WeVisDoc-2BOverall (mean of tables, charts, content faithfulness, semantic formatting, visual grounding) from leaderboard.csv in the ParseBench repo (last leaderboard commit 2026-10-09). Best of the two sizes listed (4B: 36.15). | 9 Oct 2026 | github.com |
Where you can use it
No tracked tool lists this model yet; as open weights it runs in local runtimes.