Addendum: You can try Clari yourself now on molab (molab.marimo.io/notebooks/nb...) or run it locally with pip install clari. Here's Clari generating a cocrystal with 4 caffeine and 2 oxalic acid molecules, running on an M1 MacBook Pro: x.com/auhcheng/sta...

Clari also works on non-standard chemistries like metal complexes, fullerenes, and atom clusters. Kudos to all authors: @alstonlo.bsky.social, Luka Mucko, @auhcheng.bsky.social, Andy Cai, Alastair J. A. Price, Wojciech Matusik, @aspuru.bsky.social. [4/4]

Because Clari models hydrogens explicitly, generated structures can be directly ranked with force fields such as UMA. Generating more samples (increase n_s) improves accuracy. [3/4]

Clari is built on two main design choices: 1. Model only the unit cell, rather than crops of the bulk material. 2. Replace expensive triangle layers with pair-bias attention. Input: 2D graph of each molecule in unit cell Output: unit cells in Cartesian coordinates. [2/4]

💠Introducing Clari, a generative model to predict organic crystal structure in seconds, enabling virtual screening of organic molecules on the basis of solid-state properties. 📜 arxiv.org/abs/2606.03199 👩‍💻 huggingface.co/the-matter-l... 💻 github.com/aspuru-guzik... [1/4]

Minho Kim, an undergraduate intern mentored by Jesus Valdes Hernandez, received the Grand Prize in the Marine Science and Technology Idea Competition. [4/6]

Seongmin Kim presented an invited talk, "Materealize: A Multi-Agent Deliberation System for End-to-End Material Design and Synthesis," in the AI Forum IV: Korea Institute of Industrial Technology (KITECH) Manufacturing AI session. 📜 arxiv.org/abs/2601.15743 [3/6]

Changhyeok Choi presented "El Agente Estructural: An AI Agent for Autonomous 3D Molecular Structure Generation and Editing," in the Korea Research Institute of Chemical Technology-Canada AI and Autonomous Technologies for Materials Innovation session. 📜 arxiv.org/abs/2602.04849 [2/6]

☃️ Excited to share our new paper: “Lessons From Drug Discovery for Cryoprotective Agent Design: An AI-Oriented Perspective.” 🔗 advanced.onlinelibrary.wiley.com/doi/10.1002/... [1/8]

🎉 We were honored to have our Lab Director, @aspuru.bsky.social, deliver the convocation address at this year’s University of Toronto (University College-Sciences) ceremony! 🎓 📹 www.youtube.com/live/ib2Fpfm... [1/5]

A new member of the El Agente family - “El Agente Seguro: An Agent for Chemical Safety” is now on ChemRxiv 🤖 🔗 chemrxiv.org/doi/full/10.... It focuses on one of the most important challenges for AI-driven chemistry and self-driving laboratories: safety. [1/7]

Curious to see what’s going on at the Matter Lab? Join us at CSC Toronto as our group members present their latest research in computational chemistry, machine learning, and materials science. We’d love to connect and share our work with you!

From alchemy to AI-chemy: how will AI reshape the future of chemistry and materials discovery? In Dædalus, @aspuru.bsky.social reflects on the relationship between matter, minds, and tools, from the rise of modern chemistry to today’s AI-driven systems. 🔗 www.amacad.org/publication/... [1/4]

Excited to share another work: “QCalEval: Benchmarking Vision-Language Models for Quantum Calibration Plot Understanding” 📄 arxiv.org/abs/2604.25884 🤖 huggingface.co/nvidia/Ising... 💽 huggingface.co/datasets/nvi... 📝 developer.nvidia.com/blog/nvidia-... [1/8]

🚀Excited to share our latest work: “Optimizing ground state preparation protocols with autoresearch”. We explore how language-model coding agents can automate the optimization of ground-state preparation protocols in quantum computing & classical quantum simulation. 🔗 arxiv.org/abs/2604.25610 [1/4]

Everyone is racing to build agents for science, but what how about letting scientific agents build themselves? Here we built El Agente Forjador where agents forge ⚒️ their own tools on demand for quantum chemistry and quantum dynamics. 📜 arxiv.org/abs/2604.14609

This is a platform engineering role focused on making the system reliable, scalable, and accessible to users worldwide. You’ll work alongside computational chemists, ML researchers, and HPC specialists in a leading scientific AI lab.

What if AI could write quantum circuits to help us design next-generation materials? We’re excited to share our new preprint: "Auger Spectroscopy via Generative Quantum Eigensolver: A Quantum Approach to Molecular Excitations". 📃 arxiv.org/abs/2603.12859

We’re excited to share that Kourosh Darvish and @allanzhao.bsky.social will be presenting at an upcoming event by the Canadian Physical AI Institute (CPAI), on recent advances in simulation and robot manipulation for SDLs. [1/3]

Our second agent, El Agente Estructural (“Structural” in Spanish) is a multimodal, natural-language–driven agent for molecular geometry generation and manipulation. 🔗 www.arxiv.org/abs/2602.048... [1/7]

This is achieved by granting the agent full understanding and control over ORCA (Max-Planck-Institut für Kohlenforschung) quantum chemistry software (@faccts.de), achieving remarkable handling of queries with incredible complexity and depth (see our accelerated demo below). [4/7]

⚛️ El Agente Quntur is a research collaborator agent for quantum chemistry, who can autonomously tackle difficult, long-term tasks with minimal human supervision. 🔗 www.arxiv.org/abs/2602.04850 [2/7]

Now in Digital Discovery: Context-aware computer vision for chemical reaction state detection. 🔗 pubs.rsc.org/en/content/a... [1/6]

New preprint: Materealize — a multi-agent deliberation system for end-to-end material design and synthesis. 🔗 arxiv.org/abs/2601.15743 🖥️ Demo: huggingface.co/spaces/SNU-M... [1/7]

This week on The Matter Blotter, we highlight a fascinating study on how bulky ligands influence palladium-catalyzed Suzuki-Miyaura cross-coupling, by @chertianser.bsky.social. 🔗 aspuru.substack.com/p/bulky-liga... [1/3]

New in Digital Discovery - Computer Vision Tutorial For Materials Design High-throughput synthesis has scaled. Characterization hasn’t. If you work in materials synthesis, characterization, or lab automation, this tutorial is for you. 📃 pubs.rsc.org/en/content/a... [1/6]

🧪🤖 New paper in Nature Computer Science! What if you could design, test, and refine automated chemistry workflows before touching the lab? We present “MATTERIX: toward a digital twin for robotics-assisted chemistry laboratory automation”. 🔗 www.nature.com/articles/s43... [1/6]

Happy New Year! We began the year by reading this amazing Chemistry World article by @robinson-julia.bsky.social on how AI agents are democratising computational chemistry. 🔗 www.chemistryworld.com/news/ai-agen... Image source: © Caroline Chapple, Courtesy of Chemistry World. [1/5]

We're honoured to have more than 550 subscribers since we re-launched The Matter Blotter 🤯 We have many more blog posts in the works, so subscribe here if you want to be in the loop: aspuru.substack.com

🔬 This week on The Matter Blotter, @gkwt.bsky.social presents: Ranking Models for Bayesian Optimization, a different way to think about Bayesian Optimization for self-driven molecular discovery. 👀 aspuru.substack.com/p/ranking-mo... [1/3]