HyperGVL - Order-Weighted Shortest Path: leaderboard

HyperGVL (2026) hypergraph task for vision-language models: the shortest path length between two vertices with hyperedge order as distance, with each of 200 problems posed in 35 textual and visual representations. Accuracy (%); higher is better.

Metric: Accuracy (%) on the shortest path length between two vertices with hyperedge order as distance, 7,000 questions per task (200 problems on synthetic and real-world hypergraphs, each posed in 35 textual, visual and combined hypergraph representations), accuracy (%) averaged over the representations, temperature 0.8, top-p 0.95; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 12 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 3 Flash82.09#175
2GPT-4o39.43#751
3Qwen 3 VL 8B32.1#917
4Gemma 3 12B29.41#2059
5InternVL3-8B16.23#1782
6Qwen 2.5 VL 7B Instruct15.73#1902
7GLM-4V-9B13.51#3142
8InternVL2.5-8B10.94#2370
9Molmo-7B-D-092410.53#7207
10LLaVA-NeXT-7B6.83#5032
11LLaVA-v1.5-7B5.57#7630
12DeepSeek-VL23.54#2927

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

Interactive version: theaggregate.ai/benchmark?slug=hypergvl-order-weighted-shortest-path · How It Works · Data refreshed daily, snapshot 2026-11-21.