PlantMicro - Counting MAE: leaderboard
Direct counting in PlantMicro (2026), a benchmark of vision-language models on microscopic plant images: the mean absolute error of the predicted number of target instances, zero-shot. Lower is better.
Metric: Mean absolute error of the directly predicted number of target instances (integer-only output; outputs that violate the format count as failures); zero-shot; lower is better. Source: arxiv.org. Saturation forecast: Around December 2026. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 2.07 | #167 |
| 2 | GPT-5 Mini | 2.84 | #348 |
| 3 | Gemini 2.5 Pro | 3.3 | #266 |
| 4 | Gemini 2.5 Flash | 3.39 | #453 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=plantmicro-counting-mae · How It Works · Data refreshed daily, snapshot 2026-11-13.