OWE-Bench - Sentinel-2 Gap Filling: leaderboard

OWE-Bench - Sentinel-2 Gap Filling — 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 July 2028. 14 models tracked. Updated 15 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-6 AstraMax0.4031#6
2Gemini 4 CarbonHigh0.3831#2
3Gemini 4 ArgonHigh0.3446#12
4Claude Sonnet 5.5Max0.3317#19
5GPT-6.1 SolMax0.3299#10
6Claude Fable 5.2Max0.3268#1
7Claude Opus 5.5Max0.3248#5
8Claude Fable 5.1Max0.2471#15
9Muse WonderMax0.2119#28
10GLM-5.3Max0.1676#68
11Grok 4.7xHigh0.1602#47
12DeepSeek V4.1 FlashMax0.1194#62
13Beam 1.0Max0.07351#206
14InklingMax0.007925#281

Interactive version: theaggregate.ai/benchmark?slug=owe-bench-sentinel-2-gap-filling · How It Works · Data refreshed daily, snapshot 2026-11-16.