AWARE-US (Global Weighting) - S1 Unique Repair: leaderboard
Metric: Car match (%; share of the 41 persona-grounded car-search dialogues, four preference constraints of which exactly one can be dropped to make the query feasible, in which the agent relaxes the oracle's least important constraint and recommends the oracle's car; global one-shot weighting: after the whole dialogue the model assigns normalised weights to all constraints and relaxes the lowest). Source: arxiv.org. Saturation forecast: Estimated already saturated. 3 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 8B | 24.4 |
| 2 | Gemma 2 9B (IT) | 24.4 |
| 3 | Llama 3 8B Instruct | 9.8 |
Interactive version: theaggregate.ai/benchmark?slug=aware-us-global-weighting-s1-unique-repair · How It Works · Data refreshed daily, snapshot 2026-09-26.