Imaging-101 - Function-Level: leaderboard

Metric: Function-level pass rate (%): share of the 57 expert-verified computational imaging tasks from six scientific domains whose every module passes its paired unit tests within 10 reflection rounds; higher is better. Source: arxiv.org. Saturation forecast: Around September 2028. 7 models tracked.

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)22.8
2Claude Opus 4.622.8
3GPT-5.421.1
4DeepSeek V315.8
5Qwen 3.6 Plus12.3
6Kimi K2.51.8
7GLM-51.8

Interactive version: theaggregate.ai/benchmark?slug=imaging-101-function-level · How It Works · Data refreshed daily, snapshot 2026-09-29.