XtalPi Science AI agent outperforms Claude in drug discovery, integrates robotic lab automation
XtalPi Holdings upgraded its XtalPi Science AI agent platform with over 80 proprietary scientific skills and tools, outperforming Anthropic's Claude in internal drug discovery tests. The platform now connects to autonomous laboratories for robotic experiment execution. In a molecular library synthesis project, the system improved data collection efficiency by over 40 times and reduced human expert workload by 80%, enabling a 15-person chemistry team to deliver over 20,000 reaction data points monthly. XtalPi shares rose over 5% in Hong Kong trading.
Reference imageEditorial responsibility
- No named human review is recorded for this page.
- Reports are grouped by semantic similarity and deterministic rules. Language models may assist titles, summaries, translation and cross-source analysis; the page reads the event directly, while its address stays stable when the title changes.
- Summary covers the current reports
Cross-source coverage
Reporting timeline
XtalPi Holdings Surges Over 5% as AI Platform Outperforms Claude in Drug Discovery Tasks
XtalPi Holdings (02228) rose over 5% in Hong Kong trading, reaching 8.86 HKD with a turnover of 731 million HKD. The company announced a major upgrade to its scientific AI agent platform, XtalPi Science, which now includes over 80 proprietary scientific skills and tools and has expanded its invitation-only testing. Internal tests showed the platform outperforming Anthropic's Claude model on multiple drug discovery tasks. The article notes that scientific research is becoming a key area for AI agent deployment, with Anthropic also building wet labs for robot-executed experiments. XtalPi's AI4S (AI for Science) smart solutions revenue reached 194 million yuan in the first half of the year, up 136.4% year-on-year, now accounting for 49.2% of total revenue. Guosheng Securities commented that the company has built a full-stack AI4S system integrating scientific agents, computational models, automated labs, and data feedback, and is moving from technology validation to large-scale commercialization in drug R&D and new materials.
Read sourceXtalPi Upgrades XtalPi Science Platform, Launches 80+ Proprietary Scientific Skills for AI Agents
On September 21, XtalPi Technology announced a major product upgrade to its scientific AI agent platform, XtalPi Science. The upgrade introduces over 80 proprietary scientific skills and tools, encapsulating real R&D project expertise, proprietary datasets, algorithms, and computational tools. These capabilities cover key stages of drug discovery, including target research, patent analysis, molecular generation, property prediction, protein preparation, molecular docking, retrosynthesis route prediction, structure-activity relationship analysis, LC-MS spectrum analysis, and data processing automation. The skills are presented as Tools, Skills, and Reusable Recipes, accessible via a simple dialog command without programming. The platform now extends from digital to experimental execution, connecting to a robot laboratory to schedule experiments, collect data, and adjust tasks based on feedback. XtalPi stated that as more researchers join, it will continue to drive AI, research agents, and robotic experiments to form a tighter closed loop from scientific question to analysis and experimental verification.
XtalPi Holdings Upgrades XtalPi Science Platform, Launches Over 80 Proprietary Scientific Skills
On September 21, XtalPi Holdings announced a major upgrade to its scientific agent platform, XtalPi Science, with the launch of over 80 proprietary scientific skills and tools, and an expanded invitation-only testing phase. According to the company, the platform has outperformed Claude in multiple internal drug discovery tasks. The upgrade also strengthens the platform's connection to XtalPi's autonomous laboratory, enabling it to dispatch robots to conduct experiments, collect data, and advance subsequent tasks based on experimental results. The 80+ skills and tools encapsulate the professional experience of research experts, along with proprietary datasets, algorithms, and computational tools, allowing the agent to call, combine, and trace mature research methods on demand. These capabilities cover key stages of drug discovery, including target research and patent analysis, molecular generation and property prediction, protein preparation and molecular docking, retrosynthesis route prediction, structure-activity relationship analysis, LC-MS spectrum analysis, and data processing and workflow automation.
Read sourceShow 3 older updatesHide older updates
XtalPi Holdings Upgrades XtalPi Science Platform, Launches Over 80 Proprietary Scientific Skills and Tools
On September 21, XtalPi Holdings announced a major upgrade to its scientific agent platform, XtalPi Science, which now features over 80 proprietary scientific skills and tools. The platform, which integrates expert research experience with proprietary datasets, algorithms, and computational tools, allows agents to call and combine mature research methods on demand. In internal tests, the platform outperformed Claude on multiple drug discovery tasks. The upgrade also strengthens the platform's connection to XtalPi's automated laboratories, enabling robot scheduling. The new capabilities cover key stages of drug discovery, including target research and patent analysis, molecular generation and property prediction, protein preparation and molecular docking, retrosynthetic route prediction, structure-activity relationship analysis, LC-MS spectrum interpretation, and data processing and workflow automation. The article was originally published by Shanghai Securities News.
Read sourceXtalPi Science AI Agent Outperforms Claude in Drug Discovery, Integrates Robotic Lab Automation
According to a report from Xinmin Evening News on July 21, XtalPi Technology has upgraded its scientific AI agent platform, XtalPi Science, launching over 80 proprietary scientific skills and tools, and expanding beta testing. Internal tests show the platform outperforms Anthropic's Claude in multiple drug discovery tasks. In a hit-to-lead DMTA workflow test, XtalPi Science achieved a 90% effective recommendation rate (molecules experts would advance to experiments) versus Claude's 70%, and a 60% excellent recommendation rate versus Claude's 20%. The platform also connects to XtalPi's autonomous laboratory, enabling robotic experiment execution. In a molecular library synthesis project, the system improved data collection efficiency by over 40 times, reduced human expert analysis and decision workload by 80%, and allowed a 15-person chemistry team to deliver over 20,000 reaction data points and physical molecules monthly. The article notes that AI scientific agents are increasingly entering real experimental phases, with Anthropic also building wet labs and exploring robot-scheduled experiments, validating XtalPi's long-standing AI-plus-robotics approach.
Read sourceXtalPi Science Upgrades, Outperforms Claude in Drug Discovery Tasks
On September 21, 2026, XtalPi (2228.HK) announced a major upgrade to its scientific agent platform XtalPi Science, launching over 80 proprietary scientific skills and tools. The platform outperformed Anthropic's Claude in multiple internal drug discovery tasks, including hit-to-lead design-test-synthesis-analysis workflows where XtalPi Science achieved a 90% recommendation effectiveness rate versus Claude's 70%, and a 60% excellence rate versus Claude's 20%. The upgrade extends capabilities from digital analysis to physical experimentation by connecting with XtalPi's autonomous laboratory, enabling robot scheduling for experiments and data collection. In a molecular library synthesis project, the system improved data collection efficiency by over 40 times and reduced human expert analysis workload by 80%, allowing a 15-person chemistry team to deliver over 20,000 reaction data points and physical molecules monthly. The platform covers drug discovery stages including target research, molecular generation, property prediction, retrosynthesis, and LC-MS spectrum analysis.
Read source