A 100-task expert-authored benchmark of professional charts resists saturation: the best frontier model scores 45.0%, with failures concentrated in visual perception and domain conventions.
MatCha: Enhancing visual language pretraining with math reasoning and chart derendering
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Chartography: A Benchmark for Professional Chart Understanding
A 100-task expert-authored benchmark of professional charts resists saturation: the best frontier model scores 45.0%, with failures concentrated in visual perception and domain conventions.