ERQA-Plus - Open-Ended: leaderboard
Metric: Mean Sentence-BERT cosine similarity (x100, -100 to 100) between a direct answer of at most 20 words and the reference answer over all 1,766 questions, open-ended setting; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 VL 32B (Thinking) | 61.4 |
Interactive version: theaggregate.ai/benchmark?slug=erqa-plus-open-ended · How It Works · Data refreshed daily, snapshot 2026-09-29.