PlantMicro: leaderboard
PlantMicro (2026) evaluates vision-language models on microscopic plant images from light, fluorescence and electron microscopy, with multiple-choice identification of image attributes and biological entities plus counting and localization. This board is macro accuracy in percent, the mean over the eight four-option identification tasks, zero-shot; higher is better.
Metric: Macro accuracy (%): mean of the eight identification-task accuracies; four-option multiple choice on microscopic plant images (light, fluorescence and electron microscopy) with distractors drawn from the same metadata field; zero-shot, unified prompt. Source: arxiv.org. Saturation forecast: Around June 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Flash | 63.08 | #453 |
| 2 | Gemini 2.5 Pro | 62.93 | #266 |
| 3 | Gemini 2.5 Flash Lite | 57.6 | #958 |
| 4 | GPT-5 Mini | 57.26 | #348 |
| 5 | GPT-5 | 56.79 | #167 |
| 6 | GPT-4o Mini | 51.26 | #1623 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=plantmicro · How It Works · Data refreshed daily, snapshot 2026-11-13.