Molmo-7B-D-0924: benchmark results
Provider: Allen AI. Access: Open.
Aggregate Rating 1357 ± 28, rank #7537 of 10916 rated models, from 54 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HyperGVL - Isomorphism Recognition | 45.9 | Accuracy (%) on deciding whether two hypergraphs are isomorp | 36.4 |
| HyperGVL - Order-Weighted Maximum Flow | 6.74 | Accuracy (%) on the maximum flow between two vertices with h | 36.4 |
| HyperGVL - Strict Hypercycle | 6.8 | Accuracy (%) on finding a strict hypercycle, adjacent hypere | 36.4 |
| DocHop | 16.01 | Accuracy (%) | 35 |
| HyperGVL - Understanding | 18.6 | Average accuracy (%) over the six understanding tasks (verte | 30 |
| HyperGVL - Degree-Specified Vertex Counting | 13.24 | Accuracy (%) on counting vertices with a given degree, 7,000 | 27.3 |
| HyperGVL - Hyperedge Counting | 17.94 | Accuracy (%) on counting the hyperedges of a hypergraph, 7,0 | 27.3 |
| HyperGVL - Neighbors | 7.96 | Accuracy (%) on listing the direct neighbors of a vertex, 7, | 27.3 |
| HyperGVL - Order-Filtered Neighbors | 14.4 | Accuracy (%) on listing a vertex's neighbors through hypered | 27.3 |
| HyperGVL - Order-Weighted Shortest Path | 10.53 | Accuracy (%) on the shortest path length between two vertice | 27.3 |
| HyperGVL - Reasoning | 12.83 | Average accuracy (%) over the six reasoning tasks (order-wei | 27.3 |
| ChartJudgeBench - CPA Axis Scaling | 8.5 | Pairwise judging accuracy (%; the 141 axis scaling pairs; tw | 26 |
Interactive version: theaggregate.ai/model?slug=molmo-7b-d-0924 · How It Works · Data refreshed daily, snapshot 2026-11-08.