ResearchBenchmarksVisual Reasoning on ChartFollow82.9AccuracyGPT-4o7.527.07546.6566.225May 25, 2025Evaluation ResultsMethodMethodLinksAccuracyGPT-4oInference Mode=Pure RunInference Mode=Pure Run2025.0582.9VTool-R1Model Size=7BModel Size=7B2025.0580.7GPT-4oInference Mode=Tool UseInference Mode=Tool Use2025.0580.5VTool-R1Model Size=3BModel Size=3B2025.0564R1-VLModel Size=7BModel Size=7B2025.0563.8R1-VLModel Size=2BModel Size=2B2025.0510.4