MixRea: leaderboard
Large language models (LLMs) are increasingly integrated into high-stakes decision-making. Inspired by the theory of inattentional blindness in human cognition, we investigate whether LLMs, trained on human-preferred corpora that embed attentional biases, exhibit a similar limitation: failing to attend to subtle yet important contextual cues under explicit task instructions. To eva
Metric: Implicit Information Ignorance Rate (I3R) (self-reported). Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 9.4 | #152 |
| 2 | DeepSeek R1 | 9.5 | #251 |
| 3 | Llama 3.1 405B | 12.4 | #466 |
| 4 | Claude 3.7 Sonnet | 12.7 | #249 |
| 5 | GPT-4o | 13.2 | #343 |
| 6 | Llama 3.1 70B | 13.6 | #592 |
| 7 | DeepSeek V3 | 14 | #321 |
| 8 | Gemma 2 27B | 14.3 | #806 |
| 9 | Qwen 2 7B | 14.3 | #1095 |
| 10 | Qwen 2.5 32B | 14.6 | #546 |
| 11 | Llama 3 70B | 14.6 | #631 |
| 12 | Qwen 2.5 7B | 15 | #922 |
| 13 | Gemini 2.5 Flash | 15.1 | #245 |
| 14 | Qwen Max | 15.4 | #382 |
| 15 | Qwen 2.5 72B | 16 | #499 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mixrea · How It Works · Data refreshed daily, snapshot 2026-10-20.