BBQ Ambiguous Accuracy (Sonnet 5 & Opus 5 System Cards): leaderboard
Bias Benchmark for Question Answering (BBQ) accuracy on ambiguous questions, where the correct answer is that the context is insufficient, as run by Anthropic (percent, higher is better; near the ceiling for every model). From Table 4.4.2.A of the Claude Opus 5 System Card (July 2026), with Mythos Preview and Sonnet 4.6 rows from the Sonnet 5 card.
Metric: Accuracy (%). Source: www-cdn.anthropic.com. Source: www-cdn.anthropic.com. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Mythos Preview | 100 | #5 |
| 2 | Claude Mythos 5 | 99.9 | #8 |
| 3 | Claude Opus 4.8 (Non-reasoning) | 99.9 | #39 (Claude Opus 4.8) |
| 4 | Claude Sonnet 5 (Non-reasoning) | 98.6 | #65 (Claude Sonnet 5) |
| 5 | Claude Sonnet 4.6 (Non-reasoning) | 97.5 | #94 (Claude Sonnet 4.6) |
No result here: #6 GPT-6 Astra, #7 GPT-6.1 Sol, #9 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=bbq-ambiguous-accuracy-sonnet-5-opus-5-system-cards · How It Works · Data refreshed daily, snapshot 2026-11-05.