KGCE (Without Knowledge Base): leaderboard
Multimodal LLM agents operating 104 education tasks on Windows, Android or both, many in school-specific software such as the Xiaoya assistant (KGCE, Liu et al., 2026). Completion ratio: the share of each task's sub-goals completed, in percent; higher is better. The agent has no description of the private software.
Metric: Completion ratio (%; share of task sub-goals completed; 104 educational GUI tasks on Windows, Android and cross-platform workflows with school-specific software, decomposed into DAG sub-goals; the multimodal model drives the agent; without the software knowledge base). Source: arxiv.org. Saturation forecast: Estimated already saturated. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 65.39 | #741 |
| 2 | Gemini 2.0 Flash | 61.8 | #702 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=kgce-without-knowledge-base · How It Works · Data refreshed daily, snapshot 2026-11-12.