ERR-EVAL - Missing Info: leaderboard
The "Missing Info" track of ERR-EVAL, which scores whether a model notices a defective prompt and says what it cannot answer instead of inventing it. This track withholds a fact the question needs.
Metric: Track score. Source: err-eval.gustycube.com. Saturation forecast: Not forecast. 60 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 | 6 | #115 |
| 2 | Claude Sonnet 4.6 | 5.4 | #94 |
| 3 | Claude Opus 4.6 | 5 | #67 |
| 4 | Claude Opus 4.5 | 4.6 | #87 |
| 5 | Claude Haiku 4.5 | 4.4 | #287 |
| 6 | GPT-5.4 Nano | 4 | #359 |
| 7 | GPT-5.2 Pro | 3.6 | #70 |
| 8 | Claude Sonnet 4.5 | 3.4 | #148 |
| 9 | GPT-5.4 Mini | 3.4 | #228 |
| 10 | GPT-5.1 | 3.2 | #145 |
| 11 | MiniMax-M2.5 | 3 | #316 |
| 12 | GPT-5.4 | 2.8 | #83 |
| 13 | GPT-4.1 | 2.8 | #258 |
| 14 | DeepSeek V3.2 | 2.8 | #210 |
| 15 | GLM-4.7 | 2.8 | #202 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=err-eval-missing-info · How It Works · Data refreshed daily, snapshot 2026-11-05.