ChartQA

multimodal
+
+
+
+
About

ChartQA is a benchmark for question answering about charts that requires both visual and logical reasoning capabilities. Featuring charts paired with natural language questions, it tests AI models' ability to interpret visual data representations and perform reasoning over statistical information. The benchmark challenges systems to extract insights from various chart types including bar graphs, line charts, and pie charts, essential for data analysis and visual reasoning tasks.

+
+
+
+
Evaluation Stats
Total Models24
Organizations10
Verified Results0
Self-Reported24
+
+
+
+
Benchmark Details
Max Score1
Language
en
+
+
+
+
Performance Overview
Score distribution and top performers

Score Distribution

24 models
Top Score
90.8%
Average Score
84.2%
High Performers (80%+)
20

Top Organizations

#1Anthropic
1 model
90.8%
#2Amazon
2 models
88.0%
#3Alibaba Cloud / Qwen Team
4 models
87.6%
#4Meta
4 models
86.9%
#5Mistral AI
3 models
85.8%
+
+
+
+
Leaderboard
24 models ranked by performance on ChartQA
LicenseLinks
Oct 22, 2024
Proprietary
90.8%
Apr 5, 2025
Llama 4 Community License Agreement
90.0%
Jan 26, 2025
tongyi-qianwen
89.5%
Nov 20, 2024
Proprietary
89.2%
Apr 5, 2025
Llama 4 Community License Agreement
88.8%
Aug 29, 2024
tongyi-qianwen
88.3%
Nov 18, 2024
Mistral Research License (MRL) for research; Mistral Commercial License for commercial use
88.1%
Jun 20, 2025
Apache 2.0
87.4%
Jan 26, 2025
Apache 2.0
87.3%
Nov 20, 2024
Proprietary
86.8%
Showing 1 to 10 of 24 models
+
+
+
+
Resources