We hosted our first AI x drug discovery event in NYC! A key point of discussion was that inference is getting cheaper, and that's exciting for the future of drug discovery: 1️⃣ Open weight models are closing the gap on frontier models 2️⃣ With our agent on DDBench, open-weight models matched or beat Opus 5 on 37/82 tasks and reduced costs by 70-95% in the process 3️⃣ Model selection and routing have a big impact: lower costs can unlock agent-based research at a scale that's not economical with the costs of frontier models today Thanks to everyone who came out, we all had a blast and are excited to keep building this community in NYC!