M3Exam (Multimodal Memory) - Thematic Reasoning: leaderboard

Metric: LLM-judge score (0-100) on the thematic reasoning (TH) questions of 5,150 questions over 239 long multi-session user-assistant conversations with images and PDFs (15 personas); each model answers with the annotated gold supporting sessions and their images or rendered pages placed in context, no memory module; scored by a Qwen2.5-VL-32B-Instruct judge on a five-level rubric (0, 0.25, 0.5, 0.75, 1), printed 0-1 and x100 here; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 9 models tracked.

Top models

#ModelScore
1GLM-5.161.93
2Claude Opus 4.660.83
3Qwen 3.6 Plus52.27
4Gemini 3.1 Pro (Preview)44.17
5Kimi K2.543.51
6GPT-5.443.33
7Grok 436.67
8Seed 2.0 Pro13.64
9Qwen 2.5 VL 7B Instruct6.35

Interactive version: theaggregate.ai/benchmark?slug=m3exam-multimodal-memory-thematic-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-29.