DALPHIN (Contextual) - Multiple Choice: leaderboard

Metric: Accuracy (0-1; one case-specific multiple-choice question per case with three to six options, written by the contributing pathologist; 300 questions; an unanswered question counts as wrong; contextual scenario: all questions of a case asked in sequence in one conversation). Source: arxiv.org. Saturation forecast: Around January 2028. 3 models tracked.

Top models

#ModelScore
1Gemini 2.5 Pro (Medium)0.76
2GPT-5 (Medium)0.74

Interactive version: theaggregate.ai/benchmark?slug=dalphin-contextual-multiple-choice · How It Works · Data refreshed daily, snapshot 2026-09-26.