math benchmark
ScienceQA
ScienceQA is the first large-scale multimodal science question answering benchmark with 21,208 multiple-choice questions covering 3 subjects (natural science, language science, social science), 26 topics, 127 categories, and 379 skills. The benchmark includes both text and image modalities, featuring detailed explanations and Chain-of-Thought reasoning to diagnose multi-hop reasoning ability.
Updated Aug 7, 2026
Published models1
Registry coverage1
MetricScore
EvidenceB
ScienceQA leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | MI | 91.3% | 100.0% | 1 | C |
ScienceQA highlights
The top published results on this benchmark's own scale.
What is ScienceQA?
Definition and scoring fields from the benchmark registry.
ScienceQA is the first large-scale multimodal science question answering benchmark with 21,208 multiple-choice questions covering 3 subjects (natural science, language science, social science), 26 topics, 127 categories, and 379 skills. The benchmark includes both text and image modalities, featuring detailed explanations and Chain-of-Thought reasoning to diagnose multi-hop reasoning ability.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- ScienceQA
- Modality
- multimodal
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- scienceqa|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about ScienceQA.
Which model scores highest on ScienceQA?
Phi-3.5-vision-instruct is currently ranked first with 91.3%.
What does ScienceQA measure?
ScienceQA is the first large-scale multimodal science question answering benchmark with 21,208 multiple-choice questions covering 3 subjects (natural science, language science, social science), 26 topics, 127 categories, and 379 skills. The benchmark includes both text and image modalities, featuring detailed explanations and Chain-of-Thought reasoning to diagnose multi-hop reasoning ability.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
1 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.