PICQ-drama (Multi-Task Prompting): leaderboard

Metric: Fidelity F1 (%; up to five identified personas per question matched to the annotators' missing personas by category and NLI, multi-task prompt). Source: arxiv.org. Saturation forecast: Estimated already saturated. 5 models tracked.

Top models

#ModelScore
1Qwen 3 32B48
2Llama 3.1 8B Instruct41.8
3Qwen 3 8B39.1
4Llama 3.1 70B Instruct34.6
5GPT-4.130

Interactive version: theaggregate.ai/benchmark?slug=picq-drama-multi-task-prompting · How It Works · Data refreshed daily, snapshot 2026-09-25.