ATAD - Referential Ambiguity: leaderboard
Metric: Accuracy (%) on the T5 referential ambiguity task (pick the sentence with an ambiguous referent) of ATAD's text anomaly detection problems, mean over four datasets of 100 problems per task generated by GPT-4o, Gemini-2.0-Flash, Claude-3.5-Sonnet and LLaMA-3.3-70B teacher-student-orchestrator agents; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.0 Flash Lite | 62.75 | #438 |
| 2 | GPT-3.5 Turbo | 55.75 | #849 |
| 3 | Claude 3.5 Sonnet | 53.5 | #337 |
| 4 | Claude 3 Haiku | 53.5 | #784 |
| 5 | GPT-4o Mini | 52.5 | #588 |
| 6 | Llama 3.3 70B Instruct | 52.25 | #520 |
| 7 | Gemini 2.0 Flash | 51 | #331 |
| 8 | GPT-4o | 49.25 | #333 |
| 9 | O4 Mini | 47.25 | #172 |
| 10 | Gemini 1.5 Flash | 17.5 | #619 |
| 11 | Claude 3.5 Haiku | 5.5 | #553 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=atad-referential-ambiguity · How It Works · Data refreshed daily, snapshot 2026-10-11.