BlueFin: 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; 120 held-out tasks: 75 manipulation, 9 synthesis, 36 interrogation). Source: arxiv.org. Saturation forecast: Around April 2028. 5 models tracked.

Top models

#ModelScore
1GPT-5.5 (High)49.6
2Claude Opus 4.7 (High)49.2
3Gemini 3.1 Pro (Preview) (High)45.9
4Claude Sonnet 4.6 (High)42.9

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