OWE-Bench - SmolLM3 Draft Head: leaderboard

OWE-Bench - SmolLM3 Draft Head — 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 December 2027. 14 models tracked. Updated 15 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 4 ArgonHigh0.3919#12
2Claude Fable 5.2Max0.3902#1
3Claude Sonnet 5.5Max0.3854#19
4GPT-6.1 SolMax0.3801#10
5GPT-6 AstraMax0.3691#6
6Claude Fable 5.1Max0.3641#15
7Gemini 4 CarbonHigh0.353#2
8Claude Opus 5.5Max0.3486#5
9DeepSeek V4.1 FlashMax0.1915#62
10GLM-5.3Max0.1354#68
11Muse WonderMax0.1089#28
12Grok 4.7xHigh0.06856#47
13Beam 1.0Max0.03145#206
14InklingMax0.002375#281

Interactive version: theaggregate.ai/benchmark?slug=owe-bench-smollm3-draft-head · How It Works · Data refreshed daily, snapshot 2026-11-16.