Clarify-Then-Search (1 Question): leaderboard
Metric: Nugget restoration score (0-100): mean restore_score_100, weighted nugget recall with partial credit against the static gold; one clarifying question (k=1); 518 underspecified Baidu search queries; the evaluated model asks clarifying questions, a closed-book user answerer and a rewriter (ERNIE-4.5-Turbo-128K) build the rewritten query, WebDancer searches, and ERNIE-4.5-Turbo-128K judges coverage of weighted evidence-grounded gold nuggets. Source: arxiv.org. Saturation forecast: Around 2029. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.2 | 26.47 |
| 2 | Claude Sonnet 4.5 | 25.91 |
| 3 | Gemini 2.5 Pro | 25.87 |
| 4 | DeepSeek V3.2 | 23.23 |
| 5 | Kimi K2 | 22.69 |
| 6 | Qwen 3 235B A22B Instruct | 22.53 |
Interactive version: theaggregate.ai/benchmark?slug=clarify-then-search-1-question · How It Works · Data refreshed daily, snapshot 2026-09-29.