PlantMarkerBench - Evidence Type (Arabidopsis): leaderboard
Evidence-type classification on the balanced Arabidopsis pilot split of PlantMarkerBench (2026), which asks whether a sentence from a plant biology paper supports a gene as a cell-type marker and what kind of evidence it gives: macro-F1 over expression, localization, function, indirect and noise labels, times 100. Open-weight models are served by Ollama; higher is better.
Metric: Evidence-type macro-F1 (%): macro-averaged F1 over the five evidence labels (expression, localization, function, indirect, noise) on the balanced 600-sentence Arabidopsis pilot split, times 100; open-weight models served by Ollama with the default prompt, OpenAI models with the direct prompt, temperature 0; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 13 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 32B Instruct (Ollama) | 54.5 | |
| 2 | GPT-5.4 Mini | 51.5 | #406 |
| 3 | GPT-OSS-20B (Ollama) | 43.6 | |
| 4 | Phi3-14B (Ollama) | 42 | |
| 5 | Qwen 2.5 14B Instruct (Ollama) | 41.2 | |
| 6 | Llama 3.2 3B (Ollama) | 31.7 | |
| 7 | Llama 3.1 8B (Ollama) | 29 | |
| 8 | DeepSeek-R1-8B (Ollama) | 26.8 | |
| 9 | Qwen 2.5 3B Instruct (Ollama) | 24.1 | |
| 10 | Mistral 7B Instruct (Ollama) | 23.8 | |
| 11 | Qwen 2.5 7B Instruct (Ollama) | 22.2 | |
| 12 | Phi3-3.8B-Instruct (Ollama) | 19.9 | |
| 13 | DeepSeek-R1-1.5B (Ollama) | 7.3 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=plantmarkerbench-evidence-type-arabidopsis · How It Works · Data refreshed daily, snapshot 2026-11-21.