Anthropic launched the Claude Science Beta in July 2026. This AI workbench provides scientists with over 60 specialized skills, including literature analysis, 3D protein structure rendering, and genome browsing, along with integration of the NVIDIA BioNeMo platform.
Core Function Breakdown
Claude Science allows users to upload PDF literature directly within the interface, with the system automatically extracting key data and generating structured summaries. The 3D protein structure rendering module leverages existing molecular visualization algorithms to convert amino acid sequences into interactive models, enabling users to rotate and examine binding sites. The genome browsing feature supports rapid alignment of FASTA format files, outputting results with annotated variant sites.
After integrating NVIDIA BioNeMo, users can call upon pre-trained biomolecular models for protein folding predictions. The platform provides a code execution environment where generated Python scripts can be run directly, recording input and output hash values for each computation to enable audit trails.
Technical Implementation
At the system level, it calls NVIDIA BioNeMo's inference services via APIs, translating user instructions into molecular dynamics tasks supported by BioNeMo. The code auditing feature logs the parameters of each function call to avoid black-box computations. The resource management module limits single-user GPU quotas to prevent prolonged cluster occupancy.
Validated Use Cases
A protein research team can input a gene sequence, and the system returns a predicted structure with a confidence score within 60 seconds. Literature analysis cases show the system can extract commonly mentioned mutation sites from 10 papers and list the corresponding source passages.
Currently, the Beta version is only available to Anthropic Pro subscribers. The official disclosure indicates over 60 skills in total, covering the fields of chemistry, biology, and materials science.
Impact and Limitations
This tool lowers the entry barrier for scientific research, allowing graduate students to quickly replicate others' results. However, it relies on cloud resources and is completely unusable during network outages. While code auditing exists, it cannot verify the authenticity of user-uploaded data.
The official blog disclosed that there is currently no public data on user numbers during the Beta phase.
© 2026 Winzheng.com 赢政天下 | 转载请注明来源并附原文链接