GABench - Open-Ended QA - Prediction (Numerical-Attribute Graphs): leaderboard
Metric: Task success rate (%; share of tasks whose final answer a DeepSeek-V4-Flash judge marks consistent with the verified ground truth; prediction scenario-based open-ended graph questions (insight and prediction) with its 10 analysis tools on numerical-attribute graphs (DGraph-Fin, Pokec, NBA, Credit and Bail, 200 generated tasks per dataset and category); LLM agent in the OpenClaw harness, one run). Source: arxiv.org. Saturation forecast: Around 2032. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 235B A22B | 4.36 |
| 2 | GLM-5-Turbo | 3.22 |
| 3 | GPT-5 Mini | 2.96 |
| 4 | DeepSeek V4 Pro | 1.98 |
| 5 | Gemini 3 Flash | 1.15 |
| 6 | GLM-4.5 Flash | 0.3 |
Interactive version: theaggregate.ai/benchmark?slug=gabench-open-ended-qa-prediction-numerical-attribute-graphs · How It Works · Data refreshed daily, snapshot 2026-09-29.