OpenEval - BBQ: leaderboard
Human-Centered Eval OpenEval run of BBQ exact match, measuring bias-sensitive question answering on ambiguous and disambiguated social contexts.
Metric: Exact Match (%). Source: huggingface.co. Saturation forecast: Around January 2027. 2 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | pythia-6.9B | 39 | #1646 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=openeval-bbq · How It Works · Data refreshed daily, snapshot 2026-10-20.