ToxReason - Toxicity Prediction: leaderboard
Metric: Macro F1 (%) over liver, heart and kidney toxicity labels predicted by the model on the ToxReason test set (query chemicals with curated human CTD toxicity labels and adverse outcome pathway context, eight retrieved activation and inhibition examples from similar compounds), zero-shot, temperature 0; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek R1 Distill Llama 70B | 65.7 |
| 2 | GPT-5 | 64.7 |
| 3 | O3 | 63.8 |
| 4 | Llama 3.1 70B Instruct | 63.5 |
| 5 | GPT-4o | 63.2 |
| 6 | O4 Mini | 61.8 |
| 7 | Qwen 3 4B Instruct | 61.8 |
| 8 | GPT-5.1 | 60.1 |
| 9 | Gemma 3 27B (IT) | 60 |
| 10 | Qwen 2.5 14B Instruct | 57.4 |
| 11 | Llama 3.1 8B Instruct | 49 |
Interactive version: theaggregate.ai/benchmark?slug=toxreason-toxicity-prediction · How It Works · Data refreshed daily, snapshot 2026-10-07.