Evaluates models on scientific chart comprehension and reasoning from arXiv papers, with tool access (code execution). Tests whether tool use improves chart interpretation accuracy.
This definition has not completed editorial and provenance review. Verify its metric, version, source, and evaluation configuration upstream before using it for a model-selection decision.
Tests tool-augmented vision reasoning. Difference vs no-tools variant reveals how much tool access helps.
Tool harness differences across labs can affect comparability.
Higher is better. Accuracy on chart reasoning questions with tool access.
Max score: 100
Tool harness differences across labs can affect comparability.
| # | Model | Score |
|---|---|---|
| 1 | 93.2 | |
| 2 | 91.0 | |
| 3 | 84.7 |