multimodal benchmark
CharXiv-D
CharXiv-D is the descriptive questions subset of the CharXiv benchmark, designed to assess multimodal large language models' ability to extract basic information from scientific charts. It contains descriptive questions covering information extraction, enumeration, pattern recognition, and counting across 2,323 diverse charts from arXiv papers, all curated and verified by human experts.
Updated Aug 7, 2026
Published models16
Registry coverage16
MetricScore
EvidenceB
CharXiv-D leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | BY | 95.5% | 100.0% | 16 | C | |
| 2 | BY | 94.6% | 93.3% | 16 | C | |
| 3 | AC | 90.5% | 86.7% | 16 | C | |
| 4 | AC | 90.2% | 80.0% | 16 | C | |
| 5 | OP | 90.0% | 73.3% | 16 | C | |
| 6 | OP | 88.4% | 66.7% | 16 | C | |
| 7 | CO | 88.0% | 60.0% | 16 | C | |
| 8 | OP | 87.9% | 53.3% | 16 | C | |
| 9 | AC | 86.9% | 46.7% | 16 | C | |
| 10 | AC | 85.9% | 40.0% | 16 | C | |
| 11 | AC | 85.5% | 33.3% | 16 | C | |
| 12 | OP | 85.3% | 26.7% | 16 | C | |
| 13 | AC | 83.9% | 20.0% | 16 | C | |
| 14 | AC | 83.0% | 13.3% | 16 | C | |
| 15 | AC | 76.2% | 6.7% | 16 | C | |
| 16 | OP | 73.9% | 0.0% | 16 | C |
Score distribution
Top published rows on the benchmark's original scale.
CharXiv-D highlights
The top published results on this benchmark's own scale.
What is CharXiv-D?
Definition and scoring fields from the benchmark registry.
CharXiv-D is the descriptive questions subset of the CharXiv benchmark, designed to assess multimodal large language models' ability to extract basic information from scientific charts. It contains descriptive questions covering information extraction, enumeration, pattern recognition, and counting across 2,323 diverse charts from arXiv papers, all curated and verified by human experts.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- CharXiv-D
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- charxiv-d|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about CharXiv-D.
Which model scores highest on CharXiv-D?
Seed 2.1 Pro is currently ranked first with 95.5%.
What does CharXiv-D measure?
CharXiv-D is the descriptive questions subset of the CharXiv benchmark, designed to assess multimodal large language models' ability to extract basic information from scientific charts. It contains descriptive questions covering information extraction, enumeration, pattern recognition, and counting across 2,323 diverse charts from arXiv papers, all curated and verified by human experts.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
16 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.