CrossProfile Paper Controlled HiddenBench - R1 fixed128: leaderboard
CrossProfile measures reasoning about people and events when agents retain different amounts and kinds of conversational memory. Accuracy (%).
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 2 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.1 8B Instruct | 82.1 | #1035 |
| 2 | Qwen 2.5 7B Instruct | 81 | #862 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=crossprofile-paper-controlled-hiddenbench-r1-fixed128 · How It Works · Data refreshed daily, snapshot 2026-10-29.