TIDE Study - BBH Aggregate: leaderboard
Tree-Induced Diagnostic Evaluation: 3,000 sampled items per benchmark split evenly between development and test with a fixed seed; hosted API default decoding. Common aggregate scores only; model-selected weak leaves and diagnostic method correlations are not comparable capability boards. Qwen-Plus lacks an exact dated API generation binding and remains source-qualified. Distinct from the February TIDE trajectory paper.. Score (%).
Metric: Score (%). Source: arxiv.org. Saturation forecast: Around January 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 91.27 | #63 |
| 2 | DeepSeek V3.2 | 77.13 | #204 |
| 3 | GPT-5.4 | 73.27 | #81 |
| 4 | Grok 4.20 | 71.07 | #172 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=tide-study-bbh-aggregate · How It Works · Data refreshed daily, snapshot 2026-10-23.