LDU-Bench - Morphology Recognition: leaderboard
Fine-grained morphology labelling (normal plus 11 defect categories) of the same 1,761 lithography images in LDU-Bench (2026). Macro-F1 in percent, higher is better.
Metric: Macro-F1 (%; normal plus 11 defect morphology labels on the same 1,761 images; IC-SEM lithography and integrated-circuit review images; one zero-shot run per model at temperature 0 where exposed; invalid outputs stay in the denominator; deterministic scoring, no LLM judge). Source: arxiv.org. Saturation forecast: Around 2029. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-5V Turbo | 40.8 | #342 |
| 2 | Claude Opus 4.6 | 26.6 | #100 |
| 3 | Qwen 3.6 Plus | 24.9 | #217 |
| 4 | MiniMax-M3 | 23.9 | #265 |
| 5 | GPT-5.4 | 20.1 | #121 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=ldu-bench-morphology-recognition · How It Works · Data refreshed daily, snapshot 2026-11-12.