PEC-Home (ICL) - Dynamic Preferences Level 1: leaderboard
Metric: Execution accuracy (%; 712 smart-home commands at ellipsis level 1 of 4 (complete commands), functionally correct device operations; one in-context example per ellipsis level, no dialogue history). Source: arxiv.org. Saturation forecast: Estimated already saturated. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o | 93.26 |
| 2 | DeepSeek V3 | 92.84 |
| 3 | Gemma 2 9B (IT) | 91.85 |
| 4 | Gemma 2 27B (IT) | 91.01 |
| 5 | Qwen 2.5 72B Instruct | 88.9 |
| 6 | Mistral 7B Instruct (v0.3) | 88.76 |
| 7 | Qwen 2.5 32B Instruct | 88.2 |
| 8 | Qwen 2.5 7B Instruct | 86.94 |
| 9 | Qwen 2.5 14B Instruct | 81.88 |
| 10 | Llama 3.1 8B Instruct | 74.58 |
Interactive version: theaggregate.ai/benchmark?slug=pec-home-icl-dynamic-preferences-level-1 · How It Works · Data refreshed daily, snapshot 2026-09-26.