SupraBench - Binding Affinity MAE (CoT): leaderboard
Metric: Mean absolute error of the predicted log Ka of host-guest binding under standard aqueous conditions (2,609 records), chain-of-thought prompt (explicit reasoning before the answer); lower is better. Source: arxiv.org. Saturation forecast: Around June 2027. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash | 1.26 |
| 2 | Llama 3.1 70B Instruct | 1.83 |
| 3 | GPT-5.4 Mini | 2.04 |
| 4 | GPT-5.4 Nano | 2.16 |
| 5 | Qwen 3.5 27B | 2.44 |
| 6 | Qwen 3.5 9B | 3.66 |
| 7 | Llama 3.1 8B Instruct | 4.91 |
Interactive version: theaggregate.ai/benchmark?slug=suprabench-binding-affinity-mae-cot · How It Works · Data refreshed daily, snapshot 2026-09-29.