# On the Reliability of AI Methods in Drug Discovery: Evaluation of Boltz-2 for Structure and Binding Affinity Prediction

arXiv: https://arxiv.org/abs/2603.05532 · HTML: https://mlsanity.com/paper/2603.05532/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 3 of its 50 references that are in the map.
- First posted on arXiv 2 Mar 2026.

## Builds on (most cited first, 3 of 3)

- [Automatic chemical design using a data-driven continuous representation of molecules](https://mlsanity.com/paper/1610.02415/) (2016)
- [Augmenting Genetic Algorithms with Deep Neural Networks for Exploring the Chemical Space](https://mlsanity.com/paper/1909.11655/) (2019)
- [AI Needs Physics More Than Physics Needs AI](https://mlsanity.com/paper/2512.16344/) (2025)

## Built on by (most cited first, 2 of 2)

- [Explainable quantum-compressed machine learning for complex fluid flows](https://mlsanity.com/paper/2607.21688/) (2026)
- [A Modular Agentic Framework for Synthetically Constrained Multi-Objective Hit-to-Lead Optimization](https://mlsanity.com/paper/2608.11483/) (2026)

## Abstract (arXiv)

Despite continuing hype about the role of AI in drug discovery, no "AI-discovered drugs" have so far received regulatory approval. Here we assess one of the latest AI based tools in this domain. The ability to rapidly predict protein-ligand structures and binding affinities is pivotal for accelerating drug discovery. Boltz-2, a recently developed biomolecular foundation model, aims to bridge the gap between AI efficiency and physics-based precision through a joint "co-folding" approach. In this study, we provide an extensive evaluation of Boltz-2 using two large-scale datasets: 16,780 compounds for 3CLPro and 21,702 compounds for TNKS2. We compare Boltz-2 predicted structures with traditional docking and binding affinities with binding free energies derived from the physics-based ESMACS protocol. Structural analysis reveals significant global RMSD variations, indicating that Boltz-2 predicts multiple protein conformations and ligand binding positions rather than a single converged pose. Energetic evaluations exhibit only weak to moderate correlations across the global datasets. Furthermore, a focused analysis of the top 100 compounds yields no significant correlation between the Boltz-2 predictions and the binding free energies from fine-grained ESMACS, alongside observed saturation difference in ligand structures. Our results show that while Boltz-2 offers substantial speed for initial screening, it lacks the energetic resolution required for lead identification. These findings highlight the necessity of employing physics-based methods for the reliability and refinement of AI-derived models.