KnowU-Bench - General (Hard): leaderboard
Metric: Success rate (%) on the 24 hard general tasks (explicit instructions) in the reproducible KnowU-Bench Android emulator, noisy full-history memory setting (the complete user logs plus irrelevant history), gpt-4o as the user simulator grounded in a hidden user profile, scored by rule-based checks fused with a rubric-conditioned LLM judge; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6 | 70.8 |
| 2 | Gemini 3.1 Pro (Preview) | 66.7 |
| 3 | Seed 2.0 Pro | 62.5 |
| 4 | MAI-UI-8B | 29.2 |
| 5 | Qwen 3.5 122B A10B | 25 |
| 6 | Qwen 3.5 397B A17B | 20.8 |
| 7 | Qwen 3.5 9B | 12.5 |
Interactive version: theaggregate.ai/benchmark?slug=knowu-bench-general-hard · How It Works · Data refreshed daily, snapshot 2026-10-07.