multimodal benchmark
AI2D Leaderboard
AI2D is a dataset of 4,903 illustrative diagrams from grade school natural sciences (such as food webs, human physiology, and life cycles) with over 15,000 multiple choice questions and answers. The benchmark evaluates diagram understanding and visual reasoning capabilities, requiring models to interpret diagrammatic elements, relationships, and structure to answer questions about scientific concepts represented in visual form.
Updated Aug 17, 2026
Models33
Model coverage33
MetricScore
EvidenceB
AI2D Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAN | Score94.7% | Percentile100.0% | Participants33 | EvidenceC | Evaluated |
| Rank02 | ModelAC | Score94.4% | Percentile96.9% | Participants33 | EvidenceC | Evaluated |
| Rank03 | ModelOP | Score94.2% | Percentile93.8% | Participants33 | EvidenceC | Evaluated |
| Rank04 | ModelMA | Score93.8% | Percentile90.6% | Participants33 | EvidenceC | Evaluated |
| Rank05 | ModelAC | Score93.3% | Percentile87.5% | Participants33 | EvidenceC | Evaluated |
| Rank06 | ModelMA | Score92.9% | Percentile84.4% | Participants33 | EvidenceC | Evaluated |
| Rank07 | ModelAC | Score92.9% | Percentile81.3% | Participants33 | EvidenceC | Evaluated |
| Rank08 | ModelAC | Score92.7% | Percentile78.1% | Participants33 | EvidenceC | Evaluated |
| Rank09 | ModelAC | Score92.6% | Percentile75.0% | Participants33 | EvidenceC | Evaluated |
| Rank10 | ModelME | Score92.3% | Percentile71.9% | Participants33 | EvidenceC | Evaluated |
| Rank11 | ModelME | Score91.1% | Percentile68.8% | Participants33 | EvidenceC | Evaluated |
| Rank12 | ModelAC | Score89.7% | Percentile65.6% | Participants33 | EvidenceC | Evaluated |
| Rank13 | ModelAC | Score89.5% | Percentile62.5% | Participants33 | EvidenceC | Evaluated |
| Rank14 | ModelAC | Score89.2% | Percentile59.4% | Participants33 | EvidenceC | Evaluated |
| Rank15 | ModelAC | Score88.9% | Percentile56.3% | Participants33 | EvidenceC | Evaluated |
| Rank16 | ModelAC | Score88.4% | Percentile53.1% | Participants33 | EvidenceC | Evaluated |
| Rank17 | ModelXA | Score88.3% | Percentile50.0% | Participants33 | EvidenceC | Evaluated |
| Rank18 | ModelAC | Score86.9% | Percentile46.9% | Participants33 | EvidenceC | Evaluated |
| Rank19 | ModelAC | Score85.7% | Percentile43.8% | Participants33 | EvidenceC | Evaluated |
| Rank20 | ModelAC | Score85.0% | Percentile40.6% | Participants33 | EvidenceC | Evaluated |
| Rank21 | ModelAC | Score84.9% | Percentile37.5% | Participants33 | EvidenceC | Evaluated |
| Rank22 | ModelAC | Score84.9% | Percentile34.4% | Participants33 | EvidenceC | Evaluated |
| Rank23 | ModelGO | Score84.5% | Percentile31.3% | Participants33 | EvidenceC | Evaluated |
| Rank24 | ModelGO | Score84.2% | Percentile28.1% | Participants33 | EvidenceC | Evaluated |
| Rank25 | ModelAC | Score84.1% | Percentile25.0% | Participants33 | EvidenceC | Evaluated |
| Rank26 | ModelAC | Score83.2% | Percentile21.9% | Participants33 | EvidenceC | Evaluated |
| Rank27 | ModelMI | Score82.3% | Percentile18.8% | Participants33 | EvidenceC | Evaluated |
| Rank28 | ModelDE | Score81.4% | Percentile15.6% | Participants33 | EvidenceC | Evaluated |
| Rank29 | ModelDE | Score80.0% | Percentile12.5% | Participants33 | EvidenceC | Evaluated |
| Rank30 | ModelMI | Score78.1% | Percentile9.4% | Participants33 | EvidenceC | Evaluated |
| Rank31 | ModelCO | Score77.5% | Percentile6.3% | Participants33 | EvidenceC | Evaluated |
| Rank32 | ModelGO | Score74.8% | Percentile3.1% | Participants33 | EvidenceC | Evaluated |
| Rank33 | ModelDE | Score71.6% | Percentile0.0% | Participants33 | EvidenceC | Evaluated |
AI2D Highlights
The leading models and scores on this benchmark.
AI2D Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for AI2D
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis ai2d AI model leaderboard uses descending score in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01ANClaude 3.5 SonnetAnthropic
- Score
- 94.7%
- Speed
- Up to 101 tok/s via Bedrock
Strengths
- Ranks #1 of 33 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures AI2D, not total model capability
- 02ACQwen3.6 PlusAlibaba Cloud / Qwen Team
- Score
- 94.4%
- Price
- $0.50 input / $3.0 output per 1M tokens
- Speed
- Up to 16 tok/s via Together
Strengths
- Ranks #2 of 33 compared models
- 97th percentile on this benchmark
- C evidence result
Considerations
- This result measures AI2D, not total model capability
- 03OPGPT-4oOpenAI
- Score
- 94.2%
- Price
- $2.5 input / $10 output per 1M tokens
- Speed
- Up to 132 tok/s via OpenAI
Strengths
- Ranks #3 of 33 compared models
- 94th percentile on this benchmark
- C evidence result
Considerations
- This result measures AI2D, not total model capability
- 04MAPixtral LargeMistral AI
- Score
- 93.8%
- Speed
- Up to 0.10 tok/s via Mistral AI
Strengths
- Ranks #4 of 33 compared models
- 91th percentile on this benchmark
- C evidence result
Considerations
- This result measures AI2D, not total model capability
- 05ACQwen3.5-122B-A10BAlibaba Cloud / Qwen Team
- Score
- 93.3%
- Price
- $0.40 input / $3.2 output per 1M tokens
Strengths
- Ranks #5 of 33 compared models
- 88th percentile on this benchmark
- C evidence result
Considerations
- This result measures AI2D, not total model capability
Selection summary
Best AI Models for AI2D
Claude 3.5 Sonnet currently leads AI2D with 94.7%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is AI2D?
What AI2D measures and how its scores work.
AI2D is a dataset of 4,903 illustrative diagrams from grade school natural sciences (such as food webs, human physiology, and life cycles) with over 15,000 multiple choice questions and answers. The benchmark evaluates diagram understanding and visual reasoning capabilities, requiring models to interpret diagrammatic elements, relationships, and structure to answer questions about scientific concepts represented in visual form.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- AI2D
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- ai2d|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about AI2D.
Which model scores highest on AI2D?
Claude 3.5 Sonnet is currently ranked first with 94.7%.
What does AI2D measure?
AI2D is a dataset of 4,903 illustrative diagrams from grade school natural sciences (such as food webs, human physiology, and life cycles) with over 15,000 multiple choice questions and answers. The benchmark evaluates diagram understanding and visual reasoning capabilities, requiring models to interpret diagrammatic elements, relationships, and structure to answer questions about scientific concepts represented in visual form.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
33 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.