Variational AI announced Enki 4, a major update to its generative AI platform for small-molecule drug discovery, expanding target coverage from 592 to 760 drug targets and adding support for new drug modalities like degraders and antibody drug conjugates (ADCs). The Vancouver-based company, founded by machine learning researchers from MIT, Caltech, Google Research, and Microsoft Research, says the upgrade delivers faster performance and broader application across more than 12 target classes including GPCRs, kinases, ion channels, and proteases .

What Makes Enki 4 Different From Previous Versions?

Enki 4 represents a fundamental redesign of Variational AI's generative AI platform. Rather than incremental tweaks, the company re-architected the underlying algorithms and platform design to improve both speed and accuracy. The 28% increase in target coverage, from 592 to 760 pre-trained drug targets, signals a major expansion in the types of diseases and biological pathways the AI can now tackle .

The platform's new capabilities extend beyond traditional small-molecule drugs. Enki 4 now supports the design of degraders, PROTACs (proteolysis-targeting chimeras), glues, and novel payloads for antibody drug conjugates. These are emerging drug modalities that represent some of the most innovative approaches in modern pharmaceutical development. For biotech companies, this means the AI can now help design molecules across a much wider range of therapeutic strategies .

How Does Enki 4 Speed Up the Drug Discovery Process?

• Skips Early Bottlenecks: Enki 4 aims to generate lead-like molecules ready for optimization, potentially eliminating the hit identification and hit-to-lead phases that traditionally consume months or years of research.
• Expands Target Classes: The platform now covers more than 12 target classes including GPCRs, kinases, ion channels, proteases, oxidoreductases, and hydrolases, giving researchers options across diverse disease areas.
• Improves Synthesizability: The AI generates not just potent and selective molecules, but ones that are actually feasible to synthesize in a laboratory, reducing the gap between computational design and practical chemistry.

"Enki 4 is a massive step forward for our platform. We've re-architected Enki and improved the underlying algorithms to expand target and modality coverage, while operating faster to deliver even better performance for our partners," said Ali Saberali, Co-Founder and Head of Platform at Variational AI.

Ali Saberali, Co-Founder and Head of Platform at Variational AI

Why Should Biotech Companies Care About This Update?

Drug discovery is notoriously slow and expensive. Traditional approaches to finding new drug candidates can take years and cost hundreds of millions of dollars. By generating novel, potent, and selective lead-like structures directly, Enki 4 could compress timelines significantly. The platform's ability to design molecules that are both effective and synthesizable addresses a real pain point in the industry, where computationally designed drugs sometimes prove impossible or impractical to manufacture .

The expansion to 760 targets also matters because it increases the likelihood that a biotech partner working on a specific disease will find the AI useful. With nearly 30% more targets covered than the previous version, the platform becomes relevant to a broader range of therapeutic programs .

"This Enki release continues our process of continuous model innovation. Enki 4 enables our biotech partners to start with novel leads they would not have discovered using traditional methods. Our goal is to eliminate hit ID and hit-to-lead, generating molecules ready to go directly into lead optimization," stated Handol Kim, Co-Founder and CEO at Variational AI.

Handol Kim, Co-Founder and CEO at Variational AI

What's the Bigger Picture for AI in Drug Discovery?

Enki 4 reflects a maturing trend in AI-assisted drug discovery. Rather than replacing human chemists and biologists, these tools are designed to augment their work by handling the computationally intensive task of exploring vast chemical spaces. The platform's focus on generating molecules that are not just theoretically sound but practically synthesizable shows that AI drug discovery is moving beyond academic proof-of-concept toward real-world applicability .

The addition of support for emerging modalities like PROTACs and ADCs is particularly significant. These represent some of the most cutting-edge approaches in modern drug development, and having AI assistance in designing them could accelerate innovation in areas where traditional methods are still catching up. For researchers working on difficult-to-drug targets or rare diseases, tools like Enki 4 may open doors that were previously closed .

Variational AI's continuous model innovation approach suggests that this won't be the final version. The company plans to release additional targets beyond the current 760, indicating that the expansion of Enki's capabilities is ongoing. For biotech partners and researchers, this means the tool will likely become more useful over time as it covers more disease areas and therapeutic modalities .