TRUE (Text Recognition and Understanding)
Models recognize text in images and answer questions about its content.
The publisher reports model results.
- Made by
- BAAI FlagEval
- Kind
- Documents, charts and text in images
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Items from
- Another benchmark, Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Vision
- Strongest model tested
-
Gemini 2.5 Pro Mar 2025 #104 on the capability index
- Last model added
- March 2025
- Scores refreshed
- November 20th, 2026
Gemini 2.5 Pro leads with 81.6 among models released by March 2025
Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.
The newest model with a score here came out in March 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
- 81.6
Gemini 2.5 Pro
- 75.0
Gemini 2.0 Flash
- 75.0
Gemini 2.0 Pro
- 66.9 GPT-4o
- 64.9
Qwen2.5-VL-72B
- 60.9
InternVL2_5-78B
- 59.9
Qwen2.5-VL-32B
- 57.1
Pixtral Large
- 51.7 Claude 3.7 Sonnet
- 51.2
Mistral Small 3.1
- 49.4
Qwen2.5-VL-7B
- 47.9
Molmo-72B
- 41.5
Llama 3.2 90B
- 41.0
Pixtral 12B
- 40.7 MiniCPM-o-2_6
- 38.0
InternVL2.5-8B
- 32.8
Molmo-7B-D
- 32.7
Qwen2.5-VL-3B
- 31.8 LLaVA-OneVision 72B
- 20.4
LLaVA-OneVision 7B
- 16.6
Idefics3-8B-Llama3
- 15.9
Llama 3.2 11B
- 15.4
Phi-4-multimodal
- 1
Gemini 2.5 Pro Mar 2025 81.6
- 2
Gemini 2.0 Flash Dec 2024 75.0
- 3
Gemini 2.0 Pro 75.0
- 4 GPT-4o Nov 2024 66.9
- 5
Qwen2.5-VL-72B 64.9
- 6
InternVL2_5-78B 60.9
- 7
Qwen2.5-VL-32B 59.9
- 8
Pixtral Large 57.1
- 9 Claude 3.7 Sonnet 51.7
- 10
Mistral Small 3.1 51.2
- 11
Qwen2.5-VL-7B 49.4
- 12
Molmo-72B 47.9
- 13
Llama 3.2 90B 41.5
- 14
Pixtral 12B 41.0
- 15 MiniCPM-o-2_6 40.7
- 16
InternVL2.5-8B 38.0
- 17
Molmo-7B-D 32.8
- 18
Qwen2.5-VL-3B 32.7
- 19 LLaVA-OneVision 72B 31.8
- 20
LLaVA-OneVision 7B 20.4
- 21
Idefics3-8B-Llama3 16.6
- 22
Llama 3.2 11B 15.9
- 23
Phi-4-multimodal 15.4
Source: Official TRUE (Text Recognition and Understanding) leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 81.6 | Mar 2025 | Official leaderboard | |
| 2 | | 75.0 | Dec 2024 | Official leaderboard | |
| 3 | | 75.0 | Feb 2025 | Official leaderboard | |
| 4 | GPT-4o Nov 2024 | 66.9 | Nov 2024 | Official leaderboard | |
| 5 | | 64.9 | Jan 2025 | Official leaderboard | |
| 6 | | 60.9 | Dec 2024 | Official leaderboard | |
| 7 | | 59.9 | Mar 2025 | Official leaderboard | |
| 8 | | 57.1 | Nov 2024 | Official leaderboard | |
| 9 | Claude 3.7 Sonnet | 51.7 | Feb 2025 | Official leaderboard | |
| 10 | | 51.2 | Mar 2025 | Official leaderboard | |
| 11 | | 49.4 | Jan 2025 | Official leaderboard | |
| 12 | | 47.9 | Sep 2024 | Official leaderboard | |
| 13 | | 41.5 | Sep 2024 | Official leaderboard | |
| 14 | | 41.0 | Sep 2024 | Official leaderboard | |
| 15 | MiniCPM-o-2_6 | 40.7 | Jan 2025 | Official leaderboard | |
| 16 | | 38.0 | Dec 2024 | Official leaderboard | |
| 17 | | 32.8 | Sep 2024 | Official leaderboard | |
| 18 | | 32.7 | Jan 2025 | Official leaderboard | |
| 19 | LLaVA-OneVision 72B | 31.8 | Aug 2024 | Official leaderboard | |
| 20 | | 20.4 | Aug 2024 | Official leaderboard | |
| 21 | | 16.6 | Aug 2024 | Official leaderboard | |
| 22 | | 15.9 | Sep 2024 | Official leaderboard | |
| 23 | | 15.4 | Feb 2025 | Official leaderboard |
Rankings by part
We read 7 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as HW, SceneOCR and FakeBrands. Gemini 2.5 Pro leads 3 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.
HW
23 models we track have a score here.
- 1
Gemini 2.0 Flash Dec 2024 ↑1 89.5
- 2
Gemini 2.5 Pro Mar 2025 ↓1 88.2
- 3
Qwen2.5-VL-72B ↑2 76.3
- 4
Gemini 2.0 Pro ↓1 73.7
- 5
Mistral Small 3.1 ↑5 73.7
- 6 GPT-4o Nov 2024 ↓2 60.5
- 7 Claude 3.7 Sonnet ↑2 60.5
- 8
Qwen2.5-VL-3B ↑10 57.9
- 9
Qwen2.5-VL-32B ↓2 57.9
- 10 MiniCPM-o-2_6 ↑5 55.3
SceneOCR
23 models we track have a score here.
- 1
Gemini 2.0 Pro ↑2 73.7
- 2 GPT-4o Nov 2024 ↑2 70.2
- 3
Qwen2.5-VL-72B ↑2 70.2
- 4
Gemini 2.5 Pro Mar 2025 ↓3 68.4
- 5
Qwen2.5-VL-7B ↑6 56.1
- 6
Llama 3.2 90B ↑7 54.4
- 7
Gemini 2.0 Flash Dec 2024 ↓5 54.4
- 8
Qwen2.5-VL-32B ↓1 49.1
- 9
Pixtral Large ↓1 45.6
- 10
InternVL2_5-78B ↓4 45.6
FakeBrands
23 models we track have a score here.
- 1
Gemini 2.0 Pro ↑2 75.0
- 2
Gemini 2.0 Flash Dec 2024 67.9
- 3
Gemini 2.5 Pro Mar 2025 ↓2 67.9
- 4
Qwen2.5-VL-7B ↑7 60.7
- 5 GPT-4o Nov 2024 ↓1 57.1
- 6
Qwen2.5-VL-72B ↓1 57.1
- 7
InternVL2_5-78B ↓1 50.0
- 8
Pixtral Large 42.9
- 9
Molmo-72B ↑3 35.7
- 10
Molmo-7B-D ↑7 35.7
BookVQA
23 models we track have a score here.
- 1
Gemini 2.5 Pro Mar 2025 95.4
- 2
Gemini 2.0 Flash Dec 2024 91.1
- 3
Gemini 2.0 Pro 82.9
- 4 GPT-4o Nov 2024 78.7
- 5
InternVL2_5-78B ↑1 72.4
- 6
Qwen2.5-VL-72B ↓1 55.8
- 7
Llama 3.2 90B ↑6 55.7
- 8
Pixtral Large 53.5
- 9
Molmo-72B ↑3 51.9
- 10
Qwen2.5-VL-7B ↑1 50.0
DietVQA
23 models we track have a score here.
- 1
Pixtral Large ↑7 73.2
- 2
Gemini 2.0 Pro ↑1 70.5
- 3
Qwen2.5-VL-72B ↑2 68.8
- 4
Gemini 2.5 Pro Mar 2025 ↓3 68.8
- 5
Gemini 2.0 Flash Dec 2024 ↓3 65.2
- 6
Qwen2.5-VL-32B ↑1 65.2
- 7
InternVL2_5-78B ↓1 63.4
- 8 LLaVA-OneVision 72B ↑11 62.5
- 9 Claude 3.7 Sonnet 61.6
- 10 GPT-4o Nov 2024 ↓6 60.7
Chart&Info
23 models we track have a score here.
- 1
Gemini 2.5 Pro Mar 2025 83.6
- 2
Qwen2.5-VL-32B ↑5 80.8
- 3 Claude 3.7 Sonnet ↑6 75.3
- 4
Pixtral Large ↑4 72.6
- 5
Gemini 2.0 Flash Dec 2024 ↓3 72.6
- 6
Qwen2.5-VL-72B ↓1 72.6
- 7
Gemini 2.0 Pro ↓4 72.6
- 8
Mistral Small 3.1 ↑2 72.6
- 9
Pixtral 12B ↑5 67.1
- 10 MiniCPM-o-2_6 ↑5 58.9
SceneVQA
23 models we track have a score here.
- 1
Gemini 2.5 Pro Mar 2025 80.3
- 2
Qwen2.5-VL-32B ↑5 73.2
- 3
Gemini 2.0 Flash Dec 2024 ↓1 69.0
- 4
Gemini 2.0 Pro ↓1 69.0
- 5
Qwen2.5-VL-72B 66.2
- 6 GPT-4o Nov 2024 ↓2 64.8
- 7 Claude 3.7 Sonnet ↑2 62.0
- 8
Qwen2.5-VL-7B ↑3 57.8
- 9
InternVL2_5-78B ↓3 54.9
- 10
Molmo-72B ↑2 50.7
9 models have held number one on the official leaderboard since 2023
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official TRUE (Text Recognition and Understanding) 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 TRUE (Text Recognition and Understanding) 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 20 months ago, in March 2025, for Gemini 2.5 Pro. 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 (23 entries).
- Pass Models scored 23 models we track.
- Warning Newest score The newest came 20 months ago, in March 2025, for Gemini 2.5 Pro.
- Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 2.5 Pro (#104).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on flageval-baai.github.io: 23 rows for models we track 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.