AncientDoc
Models transcribe historical Chinese pages, translate passages into modern Chinese, and answer questions about meaning, knowledge, and literary style.
The project site scores 17 models on page transcription. GJ.cool independently checked 2,580 reference texts and documented omissions and reading order errors.
- Made by
- Fudan University and ByteDance
- Released
- August 2025
- Kind
- Documents, charts and text in images
- Level
- University
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Language
- Chinese
- Interaction
- Single turn
- Items from
- Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
- Last model added
- April 2025
- Scores refreshed
- November 20th, 2026
Gemini2.5-Pro (snapshot unspecified) tops the leaderboard with 6.37 Qwen-VL-Max leads with 5.77 among models released by April 2025
Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is Qwen-VL-Max, at 5.77. This view shows only models we track, each at its best.
The newest model with a score here came out in April 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 6.37 Gemini2.5-Pro (snapshot unspecified)
- 6.16 Qwen3-VL-8B
- 5.79 Doubao-V2
- 5.77 Qwen-VL-Max
- 5.72
Qwen2.5-VL-72B
- 5.70
Qwen2.5-VL-32B
- 5.52 Qwen3-VL-4B
- 5.02
Qwen2.5-VL-7B
- 4.47 GPT-4o (snapshot unspecified)
- 4.02 Qwen2-VL-7B-Instruct
- 3.97
Qwen2.5-Omni-7B
- 3.95
InternVL3-78B
- 3.68
InternVL3-38B
- 3.68
Qwen2.5-VL-3B
- 3.63
InternVL3-8B
- 3.60
InternVL2.5-8B
- 3.51 Qwen2.5-Omni-3B
- 3.22
InternVL2.5-4B
- 3.03
Qwen2-VL-2B-Instruct
- 2.80 LLaVA-OneVision 72B
- 2.67
InternVL3-2B
- 2.07
InternVL2.5-2B
- 1.66
InternVL2.5-1B
- 1.61
InternVL3-1B
- 0.67 LLaVA1.5-7B
- 0.38 LLaVA-Onevision-5B
- 1 Gemini2.5-Pro (snapshot unspecified) 6.37
- 2 Qwen3-VL-8B 6.16
- 3 Doubao-V2 5.79
- 4 Qwen-VL-Max 5.77
- 5
Qwen2.5-VL-72B 5.72
- 6
Qwen2.5-VL-32B 5.70
- 7 Qwen3-VL-4B 5.52
- 8
Qwen2.5-VL-7B 5.02
- 9 GPT-4o (snapshot unspecified) 4.47
- 10 Qwen2-VL-7B-Instruct 4.02
- 11
Qwen2.5-Omni-7B 3.97
- 12
InternVL3-78B 3.95
- 13
InternVL3-38B 3.68
- 14
Qwen2.5-VL-3B 3.68
- 15
InternVL3-8B 3.63
- 16
InternVL2.5-8B 3.60
- 17 Qwen2.5-Omni-3B 3.51
- 18
InternVL2.5-4B 3.22
- 19
Qwen2-VL-2B-Instruct 3.03
- 20 LLaVA-OneVision 72B 2.80
- 21
InternVL3-2B 2.67
- 22
InternVL2.5-2B 2.07
- 23
InternVL2.5-1B 1.66
- 24
InternVL3-1B 1.61
- 25 LLaVA1.5-7B 0.67
- 26 LLaVA-Onevision-5B 0.38
Source: Official AncientDoc leaderboard
- 5.77 Qwen-VL-Max
- 5.72
Qwen2.5-VL-72B
- 5.70
Qwen2.5-VL-32B
- 5.02
Qwen2.5-VL-7B
- 4.02 Qwen2-VL-7B-Instruct
- 3.97
Qwen2.5-Omni-7B
- 3.95
InternVL3-78B
- 3.68
InternVL3-38B
- 3.68
Qwen2.5-VL-3B
- 3.63
InternVL3-8B
- 3.60
InternVL2.5-8B
- 3.22
InternVL2.5-4B
- 3.03
Qwen2-VL-2B-Instruct
- 2.80 LLaVA-OneVision 72B
- 2.67
InternVL3-2B
- 2.07
InternVL2.5-2B
- 1.66
InternVL2.5-1B
- 1.61
InternVL3-1B
- 1 Qwen-VL-Max 5.77
- 2
Qwen2.5-VL-72B 5.72
- 3
Qwen2.5-VL-32B 5.70
- 4
Qwen2.5-VL-7B 5.02
- 5 Qwen2-VL-7B-Instruct 4.02
- 6
Qwen2.5-Omni-7B 3.97
- 7
InternVL3-78B 3.95
- 8
InternVL3-38B 3.68
- 9
Qwen2.5-VL-3B 3.68
- 10
InternVL3-8B 3.63
- 11
InternVL2.5-8B 3.60
- 12
InternVL2.5-4B 3.22
- 13
Qwen2-VL-2B-Instruct 3.03
- 14 LLaVA-OneVision 72B 2.80
- 15
InternVL3-2B 2.67
- 16
InternVL2.5-2B 2.07
- 17
InternVL2.5-1B 1.66
- 18
InternVL3-1B 1.61
Source: Official AncientDoc leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Gemini2.5-Pro (snapshot unspecified) | 6.37 | Official leaderboard | ||
| 2 | Qwen3-VL-8B | 6.16 | Official leaderboard | ||
| 3 | Doubao-V2 | 5.79 | Official leaderboard | ||
| 4 | Qwen-VL-Max | 5.77 | Jan 2024 | Official leaderboard | |
| 5 | | 5.72 | Jan 2025 | Official leaderboard | |
| 6 | | 5.70 | Mar 2025 | Official leaderboard | |
| 7 | Qwen3-VL-4B | 5.52 | Official leaderboard | ||
| 8 | | 5.02 | Jan 2025 | Official leaderboard | |
| 9 | GPT-4o (snapshot unspecified) | 4.47 | Official leaderboard | ||
| 10 | Qwen2-VL-7B-Instruct | 4.02 | Aug 2024 | Official leaderboard | |
| 11 | | 3.97 | Mar 2025 | Official leaderboard | |
| 12 | | 3.95 | Apr 2025 | Official leaderboard | |
| 13 | | 3.68 | Apr 2025 | Official leaderboard | |
| 14 | | 3.68 | Jan 2025 | Official leaderboard | |
| 15 | | 3.63 | Apr 2025 | Official leaderboard | |
| 16 | | 3.60 | Dec 2024 | Official leaderboard | |
| 17 | Qwen2.5-Omni-3B | 3.51 | Official leaderboard | ||
| 18 | | 3.22 | Dec 2024 | Official leaderboard | |
| 19 | | 3.03 | Aug 2024 | Official leaderboard | |
| 20 | LLaVA-OneVision 72B | 2.80 | Aug 2024 | Official leaderboard | |
| 21 | | 2.67 | Apr 2025 | Official leaderboard | |
| 22 | | 2.07 | Dec 2024 | Official leaderboard | |
| 23 | | 1.66 | Dec 2024 | Official leaderboard | |
| 24 | | 1.61 | Apr 2025 | Official leaderboard | |
| 25 | LLaVA1.5-7B | 0.67 | Official leaderboard | ||
| 26 | LLaVA-Onevision-5B | 0.38 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Qwen-VL-Max | 5.77 | Jan 2024 | Official leaderboard | |
| 2 | | 5.72 | Jan 2025 | Official leaderboard | |
| 3 | | 5.70 | Mar 2025 | Official leaderboard | |
| 4 | | 5.02 | Jan 2025 | Official leaderboard | |
| 5 | Qwen2-VL-7B-Instruct | 4.02 | Aug 2024 | Official leaderboard | |
| 6 | | 3.97 | Mar 2025 | Official leaderboard | |
| 7 | | 3.95 | Apr 2025 | Official leaderboard | |
| 8 | | 3.68 | Apr 2025 | Official leaderboard | |
| 9 | | 3.68 | Jan 2025 | Official leaderboard | |
| 10 | | 3.63 | Apr 2025 | Official leaderboard | |
| 11 | | 3.60 | Dec 2024 | Official leaderboard | |
| 12 | | 3.22 | Dec 2024 | Official leaderboard | |
| 13 | | 3.03 | Aug 2024 | Official leaderboard | |
| 14 | LLaVA-OneVision 72B | 2.80 | Aug 2024 | Official leaderboard | |
| 15 | | 2.67 | Apr 2025 | Official leaderboard | |
| 16 | | 2.07 | Dec 2024 | Official leaderboard | |
| 17 | | 1.66 | Dec 2024 | Official leaderboard | |
| 18 | | 1.61 | Apr 2025 | Official leaderboard |
Rankings by part
We read 5 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Page-level OCR, Vernacular Translation and Reasoning-based QA. Qwen-VL-Max leads 2 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Page-level OCR
18 models we track have a score here.
- 1
Qwen2.5-VL-7B ↑3 6.37
- 2
Qwen2.5-VL-72B 5.73
- 3 Qwen-VL-Max ↓2 5.70
- 4
Qwen2.5-VL-32B ↓1 5.67
- 5
Qwen2.5-VL-3B ↑4 5.56
- 6
InternVL3-8B ↑4 5.06
- 7 Qwen2-VL-7B-Instruct ↓2 4.66
- 8
Qwen2.5-Omni-7B ↓2 4.30
- 9
InternVL3-78B ↓2 4.04
- 10
Qwen2-VL-2B-Instruct ↑3 3.79
Vernacular Translation
18 models we track have a score here.
- 1 Qwen-VL-Max 3.22
- 2
Qwen2.5-VL-72B 2.98
- 3
Qwen2.5-VL-32B 2.87
- 4
Qwen2.5-VL-7B 2.30
- 5
InternVL3-78B ↑2 1.21
- 6 Qwen2-VL-7B-Instruct ↓1 1.17
- 7
Qwen2.5-VL-3B ↑2 0.66
- 8
InternVL2.5-8B ↑3 0.58
- 9
InternVL3-38B ↓1 0.56
- 10
InternVL2.5-4B ↑2 0.53
Reasoning-based QA
18 models we track have a score here.
- 1
Qwen2.5-VL-32B ↑2 7.49
- 2 Qwen-VL-Max ↓1 7.46
- 3
Qwen2.5-VL-72B ↓1 7.43
- 4
Qwen2.5-VL-7B 6.44
- 5
Qwen2.5-Omni-7B ↑1 6.05
- 6
InternVL2.5-8B ↑5 5.93
- 7 Qwen2-VL-7B-Instruct ↓2 5.79
- 8
InternVL3-38B 5.78
- 9 LLaVA-OneVision 72B ↑5 5.48
- 10
InternVL3-78B ↓3 5.18
Knowledge-based QA
18 models we track have a score here.
- 1
Qwen2.5-VL-72B ↑1 6.84
- 2
Qwen2.5-VL-32B ↑1 6.84
- 3 Qwen-VL-Max ↓2 6.78
- 4
Qwen2.5-VL-7B 5.23
- 5
InternVL3-78B ↑2 5.18
- 6
InternVL3-38B ↑2 5.11
- 7 LLaVA-OneVision 72B ↑7 5.04
- 8
Qwen2.5-Omni-7B ↓2 4.94
- 9
InternVL2.5-8B ↑2 4.88
- 10 Qwen2-VL-7B-Instruct ↓5 4.83
Linguistic Variant QA
18 models we track have a score here.
- 1 Qwen-VL-Max 5.67
- 2
Qwen2.5-VL-72B 5.63
- 3
Qwen2.5-VL-32B 5.61
- 4
Qwen2.5-VL-7B 4.75
- 5
InternVL3-78B ↑2 4.13
- 6
Qwen2.5-Omni-7B 4.12
- 7
InternVL3-38B ↑1 3.99
- 8
Qwen2.5-VL-3B ↑1 3.75
- 9 Qwen2-VL-7B-Instruct ↓4 3.67
- 10
InternVL2.5-8B ↑1 3.63
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official AncientDoc leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official AncientDoc leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Vision
- MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy.
Qwen 3.6 Plus 86%
- MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. GPT-5.6 Sol 84.6%
- CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
- MathVista In use Math reasoning in visual contexts (charts, geometry, puzzles) from 31 datasets. Answer accuracy.
Seed2.1 Pro 90.7%
- MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy.
Qwen 3.8 Max 97.7%
- MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy.
Qwen 3.8 Max 85.9%
Data audit
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 19 months ago, in April 2025, for InternVL3-1B. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (26 entries).
- Pass Models scored 18 models we track, plus 8 other entries on its leaderboard.
- Warning Newest score The newest came 19 months ago, in April 2025, for InternVL3-1B.
- Note Frontier models None of its models are on the capabilities index.
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on bytedance.github.io: 18 rows for models we track and 8 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.