PPT-Bench (Epistemic Attack) - Epistemic Destabilization: leaderboard

Capitulation on the 23 epistemic destabilization items of PPT-Bench (2026), philosophical pressure drawn from Pyrrhonism, Madhyamaka and Academic Skepticism; a GPT-4o judge compares each pressured answer with the unpressured baseline answer of the same model. Share of items with partial or full capitulation, in percent; lower is better.

Metric: Binary capitulation rate (%): share of items scored 1 or 2 over the 23 Type 1 items (epistemic destabilization: Pyrrhonism, Madhyamaka, Academic Skepticism); a gpt-4o pairwise judge compares each single-turn pressured answer (L1) with the model's own unpressured baseline answer (L0) and scores 0 (held), 1 (partial) or 2 (full capitulation); temperature 0, one generation per item; lower is better. Source: arxiv.org. Saturation forecast: Around February 2027. 7 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1DeepSeek V3.130.4#511
2Nemotron 3 Super34.8#773
3GPT-OSS-20BLow39.1#1183
4Llama 3.3 70B Instruct65.2#1385
5Qwen 3 32B65.2#972
6Mistral Small 3.169.6#1706
7Ministral 8B82.6#4857

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=ppt-bench-epistemic-attack-epistemic-destabilization · How It Works · Data refreshed daily, snapshot 2026-11-21.