PRISM-BN - State F1: leaderboard

Metric: State F1 (%; macro-averaged over nodes matched in the node phase, so conditional on node alignment; zero-shot extraction of a parameterized Bayesian network from each of the 5,054 PRISM-BN text descriptions through the four-phase prompt pipeline; predicted node and state names are aligned one-to-one with the reference by a Llama 3.3 70B semantic judge). Source: arxiv.org. Saturation forecast: Estimated already saturated. 6 models tracked.

Top models

#ModelScore
1Claude Haiku 4.593.1
2DeepSeek V391.8
3Qwen 3 30B A3B 2507 Instruct90.7
4GPT-4o Mini90.3
5Gemma 3 12B (IT)89.3
6Llama 4 Maverick88.6

Interactive version: theaggregate.ai/benchmark?slug=prism-bn-state-f1 · How It Works · Data refreshed daily, snapshot 2026-09-26.