Extract Bench (Contextual AI)
Models extract structured JSON from PDFs. Field checks include exact matches, numerical tolerances and semantic judgments.
The suite has 35 tasks across five schemas.
- Made by
- Contextual AI
- Released
- February 2026
- Kind
- Documents, charts and text in images
- Level
- University
- Scored
- Graded by a model
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Items from
- Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- GPT-5.2 #51 on the capability index
- Last model added
- December 2025
- Scores refreshed
- November 20th, 2026
Gemini 3 Flash leads with 0.07 among models released by December 2025
Best score per model from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart.
The newest model with a score here came out in December 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
- 1
Gemini 3 Flash Baseline 0.07
- 2
Gemini 3 Pro Baseline 0.06
- 3 GPT-5.2 Structured outputs 0.05
- 4 GPT-5 Baseline 0.04
- 5 Claude Sonnet 4.5 Baseline 0.03
- 6 Claude Opus 4.5 Baseline 0.03
Source: The Extract Bench (Contextual AI) paper
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 0.07 | Baseline | Dec 2025 | Official leaderboard |
| 2 | | 0.06 | Baseline | Nov 2025 | Official leaderboard |
| 3 | GPT-5.2 | 0.05 | Structured outputs | Dec 2025 | Official leaderboard |
| 4 | GPT-5 | 0.04 | Baseline | Aug 2025 | Official leaderboard |
| 5 | Claude Sonnet 4.5 | 0.03 | Baseline | Sep 2025 | Official leaderboard |
| 6 | Claude Opus 4.5 | 0.03 | Baseline | Nov 2025 | Official leaderboard |
Rankings by part
We read 5 parts from the benchmark’s paper, each ranking models on one part of the test, such as Credit, Research and Resumes. Gemini 3 Flash 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.
Credit
6 models we track have a score here.
- 1
Gemini 3 Pro ↑1 0.89
- 2 GPT-5 ↑2 0.87
- 3
Gemini 3 Flash ↓2 0.85
- 4 GPT-5.2 ↓1 0.83
- 5 Claude Sonnet 4.5 0.00
- 6 Claude Opus 4.5 0.00
Research
6 models we track have a score here.
- 1
Gemini 3 Flash 0.54
- 2 GPT-5.2 ↑1 0.51
- 3 Claude Sonnet 4.5 ↑2 0.49
- 4 GPT-5 0.32
- 5 Claude Opus 4.5 ↑1 0.17
- 6
Gemini 3 Pro ↓4 0.09
Resumes
6 models we track have a score here.
- 1 Claude Opus 4.5 ↑5 0.24
- 2 Claude Sonnet 4.5 ↑3 0.21
- 3
Gemini 3 Pro ↓1 0.21
- 4
Gemini 3 Flash ↓3 0.21
- 5 GPT-5.2 ↓2 0.20
- 6 GPT-5 ↓2 0.04
Sports
6 models we track have a score here.
- 1
Gemini 3 Flash 0.18
- 2 Claude Sonnet 4.5 ↑3 0.17
- 3 Claude Opus 4.5 ↑3 0.17
- 4 GPT-5.2 ↓1 0.17
- 5
Gemini 3 Pro ↓3 0.13
- 6 GPT-5 ↓2 0.03
SEC 10-K/Q
6 models we track have a score here.
- 1 GPT-5 ↑3 0.00
- 2 Claude Sonnet 4.5 ↑3 0.00
- 3
Gemini 3 Pro ↓1 0.00
- 4 Claude Opus 4.5 ↑2 0.00
- 5 GPT-5.2 ↓2 0.00
- 6
Gemini 3 Flash ↓5 0.00
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The Extract Bench (Contextual AI) paper
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Baseline | 0.07 | ||||
| Baseline | 0.06 | ||||
| Structured outputs | 0.06 | ||||
| GPT-5.2 | Structured outputs | 0.05 | |||
| GPT-5.2 | Baseline | 0.05 | |||
| Structured outputs | 0.04 | ||||
| GPT-5 | Baseline | 0.04 | |||
| GPT-5 | Structured outputs | 0.04 | |||
| Claude Sonnet 4.5 | Baseline | 0.03 | |||
| Claude Opus 4.5 | Baseline | 0.03 | |||
| Claude Sonnet 4.5 | Structured outputs | 0.00 | |||
| Claude Opus 4.5 | Structured outputs | 0.00 |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The Extract Bench (Contextual AI) paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Extract Bench (Contextual AI) 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 11 months ago, in December 2025, for Gemini 3 Flash. 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 benchmark’s paper (6 entries).
- Pass Models scored 6 models we track.
- Warning Newest score The newest came 11 months ago, in December 2025, for Gemini 3 Flash.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.2 (#51).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 12 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.