Multimodal Code Generation on ChartMimic benchmark leaderboard · SOTA2 Research

Multimodal Code Generation on ChartMimic

0.978Execution RateClaude-4.5-Sonnet

Evaluation Results

Method	Links
Claude-4.5-Sonnet2026.01		0.978	0.896	0.829
Gemini-2.5-Pro2026.01		0.973	0.887	0.838
GPT-52026.01		0.948	0.819	0.783
Qwen2.5-VL-72B2026.01		0.885	0.727	0.791
OCRVerse2026.01		0.848	0.722	0.754
Qwen3-VL-32B2026.01		0.83	0.669	0.775
InternVL3.5-38B2026.01		0.79	0.6	0.718
Qwen3-VL-8B2026.01		0.783	0.625	0.678
InternVL3.5-14B2026.01		0.732	0.528	0.554
InternVL3-14B2026.01		0.723	0.513	0.541
Qwen2.5-VL-7B2026.01		0.687	0.422	0.401
InternVL3.5-8B2026.01		0.667	0.46	0.483
InternVL3-8B2026.01		0.633	0.438	0.461