GDPval — leaderboard

Real-world, economically valuable knowledge work tasks across 44 occupations and 9 U.S. GDP sectors.

Metric: Wins/Ties vs Human (self-reported). Source: benchmarklist.com. Status: saturated. 18 models tracked.

Top models

#ModelScore
1GPT-5.584.9
2GPT-5.483
3GPT-5.5 Pro82.3
4GPT-5.4 Pro (xHigh)82
5Claude Opus 4.780.3
6Gemini 3.1 Pro (Preview)67.3
7GLM-5.154.7
8DeepSeek V4 Pro54.6
9DeepSeek V4 Flash50.2
10MiniMax-M2.747.6
11Claude Opus 4.147.6
12Nemotron 3 Ultra46.7
13GPT-539
14O335.2
15Qwen 3.5 397B A17B34.6

Interactive version: theaggregate.ai/benchmark?slug=gdpval · How the rankings work · Data refreshed daily, snapshot 2026-07-22.