BlueFin - Manipulation: leaderboard

Metric: Criteria-weighted rubric score (%; points met over maximum points of expert rubrics checked by an agentic GPT-5.4 judge for synthesis and manipulation workbooks, answer correctness for interrogation, non-completions count as zero; 20-tool spreadsheet harness, high reasoning effort; 75 held-out tasks modifying an existing financial model). Source: arxiv.org. Saturation forecast: Around May 2028. 5 models tracked.

Top models

#ModelScore
1Claude Opus 4.7 (High)49.4
2GPT-5.5 (High)48.2
3Gemini 3.1 Pro (Preview) (High)46.1
4Claude Sonnet 4.6 (High)43

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