OWE-Bench - Science: leaderboard

OWE-Bench - Science — Expert-built long-horizon science, software, ML and engineering tasks with task-specific hardware and4–18hour budgets; native baseline/ceiling normalized held-out reward,mean of8attempts and equal task means. Web frozen October1,2026;deterministic sealed verifier. Native model/effort and every attempt retained. Score fraction0–1;higher is better.

Metric: Mean normalized held-out task score (fraction, 0–1). Source: openworldengineer.mercor.com. Saturation forecast: Around March 2028. 14 models tracked. Updated 15 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Claude Fable 5.2Max0.4019#1
2GPT-6 AstraMax0.3936#6
3Claude Opus 5.5Max0.3035#5
4GPT-6.1 SolMax0.2958#10
5Claude Sonnet 5.5Max0.2823#19
6Gemini 4 CarbonHigh0.2522#2
7Gemini 4 ArgonHigh0.2276#12
8Claude Fable 5.1Max0.1973#15
9Muse WonderMax0.1437#28
10Grok 4.7xHigh0.08283#47
11GLM-5.3Max0.08218#68
12DeepSeek V4.1 FlashMax0.06737#62
13Beam 1.0Max0.03702#206
14InklingMax0.01189#281

Interactive version: theaggregate.ai/benchmark?slug=owe-bench-science · How It Works · Data refreshed daily, snapshot 2026-11-16.