safety benchmark
FigQA Leaderboard
FigQA is a multiple-choice benchmark for interpreting scientific figures from biology papers.
Updated Sep 5, 2026
Models3
Model coverage3
MetricScore
EvidenceB
FigQA Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAN | Score89.00% | Percentile100.00% | Participants3 | EvidenceC | Evaluated |
| Rank02 | ModelAN | Score78.30% | Percentile50.00% | Participants3 | EvidenceC | Evaluated |
| Rank03 | ModelXA | Score34.00% | Percentile0.00% | Participants3 | EvidenceC | Evaluated |
FigQA Highlights
The leading models and scores on this benchmark.
FigQA Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for FigQA
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is FigQA?
What FigQA measures and how its scores work.
FigQA is a safety benchmark based on multiple-choice questions about scientific figures from biology papers.
It measures dual-use biological knowledge and multimodal reasoning relevant to bioweapons development using the Score metric, reported as a ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- FigQA
- Modality
- multimodal
- Primary category
- safety
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about FigQA.
Which model scores highest on FigQA?
Claude Mythos Preview is currently ranked first with 89.00%.
What are the top three models on FigQA?
The current leaders are Claude Mythos Preview (89.00%), Claude Opus 4.6 (78.30%), and Grok-4.1 Thinking (34.00%).
Which FigQA model has the lowest official input price?
Claude Opus 4.6 has the lowest matched official input price at $5.0 input / $25 output per 1M tokens.
Which models are fastest among FigQA results?
The fastest matched records are Claude Opus 4.6 (123.15 tok/s via Anthropic) and Grok-4.1 Thinking (80.00 tok/s via xAI).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.