Finding the Right Fit - ALE-CLI (OpenHands): leaderboard

ALE-CLI, 99 professional command-line workflows run in local Docker, from the Finding the Right Fit harness study: every model runs inside the OpenHands agent harness at high reasoning effort, one counted run per task. Mean task score in percent, higher is better (Li et al., 2026).

Metric: Task score (%), OpenHands harness, high reasoning effort, one run per task over 99 tasks. Source: huggingface.co. Saturation forecast: Around 2029. 5 models tracked. Updated 5 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-6 AstraHigh53.86#10
2Claude Opus 5High53.14#26
3GLM-5.3High51.68#73
4DeepSeek V4 ProHigh51.2#198
5Kimi K3High39.15#61

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=finding-the-right-fit-ale-cli-openhands · How It Works · Data refreshed daily, snapshot 2026-11-20.