DRAGON (Diagrams) - MapIQ: leaderboard

DRAGON (2026) evidence-grounded reasoning on maps from MapIQ: given the question and its answer, the model must box every diagram region needed to justify the answer. Grounding IoU (%) between predicted and human-verified evidence regions, averaged over three prompting strategies; higher is better.

Metric: Evidence grounding on maps from MapIQ: given the diagram, the question and the verified answer, the model returns bounding boxes for every visual element needed to justify the answer; Grounding IoU (%) between the union of predicted boxes and the union of the human-verified gold evidence boxes, averaged over the EDGE (direct), SAGE (select then ground) and VERGE (verify and refine) prompting strategies; DRAGON test split; temperature 0.2, top-p 0.7; higher is better. Source: arxiv.org. Saturation forecast: Around September 2027. 8 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Claude Opus 4.628.4#113
2Claude Sonnet 4.621.3#173
3Kimi K2.517.4#266
4Llama 4 Maverick Instruct6.2#1157
5Qwen 3.5 35B A3B4.7#492
6Gemma 3 27B (IT)4.5#1324
7Gemini 3 Pro (Preview)4.3#119
8InternVL3.5-38B2.5#1021

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=dragon-diagrams-mapiq · How It Works · Data refreshed daily, snapshot 2026-11-21.