Anthropic Launches Claude Science on 1 July 2026
On 30 June 2026, Anthropic unveiled Claude Science — a purpose-built AI workbench for computational researchers — and made it available to all paid subscribers in beta starting 1 July. The launch marks Anthropic's first vertical product category outside the core Claude assistant, and its direct entry into the scientific computing market alongside OpenAI's GPT Rosalind for drug discovery and Google's research-oriented Gemini deployments.
Claude Science is not a new model. It is a multi-agent application built on top of Anthropic's existing models that pre-configures the scientific context researchers would otherwise spend months building themselves: connections to more than 60 scientific databases, specialised toolkits for biological data types, and a reviewer agent that independently checks citations and calculations before output reaches the researcher.
How the Multi-Agent Architecture Works
The centrepiece of Claude Science is a generalist coordinating agent that orchestrates more than 60 curated sub-agents and database connectors pre-configured for genomics, single-cell analysis, proteomics, structural biology, and cheminformatics. When a researcher asks a question in plain English — for example, the expression profile of a gene in a specific tissue context — the coordinator dispatches sub-agents to query relevant databases such as GEO, Ensembl, UniProt, PDB, ChEMBL, Reactome, and ClinVar, synthesises the results across sources, and returns a structured answer.
A separate reviewer agent operates in parallel, verifying citations and checking calculations for errors. Outputs are flagged and corrected before the researcher sees the final response. The reviewer agent targets the specific failure mode that has made general-purpose AI tools difficult to trust in scientific settings: confident incorrect answers that pass casual inspection but fail peer scrutiny.
Claude Science renders rich scientific artefacts natively: three-dimensional protein structures, genome browser tracks, chemical structure diagrams, and other domain-specific visualisations that allow researchers to inspect data in context rather than export it to a separate rendering tool.
Drug Discovery for Neglected Diseases
Alongside the Claude Science launch, Anthropic announced its own internal pre-clinical drug discovery programme targeting neglected diseases — conditions where low commercial incentive has historically meant minimal pharmaceutical investment. The programme is run directly by Anthropic rather than through a pharmaceutical partner. This differentiates it from OpenAI's drug discovery work, which is structured around partnerships with established pharma companies. Anthropic is running its own target selection and compound evaluation pipeline, with neglected tropical diseases and conditions endemic to low-income markets as stated focus areas.
The AI for Science Grant Programme
Anthropic will fund up to 50 Claude Science projects with grants of up to $30,000 in API credits each. Applications are open through 15 July 2026, with award notifications sent by 31 July. Funded projects will run from 1 September to 1 December 2026, and are explicitly aimed at graduate and postdoctoral researchers rather than corporate teams.
Indian researchers at IITs, IISERs, and IISc are eligible to apply. The application window closes 15 July 2026 — approximately ten days from the date of this article.
What Claude Science Means for Indian Researchers and Software Teams
India's publicly funded research institutions — the IITs, IISER campuses, IISc, CCMB, and NCBS — collectively run significant computational biology and cheminformatics workloads. For research groups that have relied on individual database API integrations and custom bioinformatics pipelines assembled from publicly available tools, Claude Science's pre-configured environment offers a substantial reduction in pipeline setup time. The 60-plus database connections and specialised sub-agents address fragmentation that is particularly acute in genomics and structural biology, where data is distributed across dozens of incompatible schemas and query languages.
For Indian life sciences software companies building drug discovery tools, clinical genomics platforms, or agricultural genomics applications, Claude Science sets a new baseline for what AI-native research tooling looks like. Teams selling research software products to pharma or biotech clients globally will need to assess whether their tools complement or compete with an Anthropic-backed workbench pre-connected to the dominant scientific databases in their domain.
The Bottom Line
Anthropic launched Claude Science in beta on 1 July 2026 — a multi-agent AI workbench pre-configured for genomics, proteomics, structural biology, and cheminformatics, with connections to more than 60 scientific databases including UniProt, PDB, Ensembl, ChEMBL, Reactome, and ClinVar. A reviewer agent checks citations and calculations in real time. Available to all paid Anthropic subscribers, Claude Science ships alongside an internal drug discovery programme targeting neglected diseases and a grant programme funding up to 50 research projects with up to $30,000 in credits each — applications close 15 July 2026. For Indian research institutions and life sciences software teams, the launch raises the baseline for AI-native scientific computing and opens an immediate grant access window.
Frequently Asked Questions
What is Anthropic Claude Science and when did it launch?+
Claude Science is a multi-agent AI research workbench launched by Anthropic in public beta on 1 July 2026, available to all paid Claude subscribers. It is not a new model but a multi-agent application built on Anthropic's existing Claude models, pre-configured for genomics, single-cell analysis, proteomics, structural biology, and cheminformatics. It connects to more than 60 curated scientific databases including UniProt, PDB, Ensembl, ChEMBL, Reactome, ClinVar, and GEO. A built-in reviewer agent checks citations and calculations in real time before output is presented to the researcher.
What scientific databases does Claude Science connect to?+
Claude Science ships pre-configured with connections to more than 60 scientific databases and toolkits across multiple domains. In biology and genomics, these include GEO, Ensembl, UniProt, PDB (Protein Data Bank), Reactome, ClinVar, and ChEMBL for cheminformatics. When a researcher poses a question in plain English, a generalist coordinating agent dispatches specialised sub-agents to query the appropriate databases and synthesise results across sources, removing the need to navigate incompatible schemas and query languages individually.
What is Anthropic's drug discovery programme for neglected diseases?+
Alongside the Claude Science launch, Anthropic announced an internal pre-clinical drug discovery programme targeting neglected diseases — including neglected tropical diseases and conditions endemic to low-income markets where low commercial incentive has historically meant minimal pharmaceutical investment. Unlike OpenAI's drug discovery work, which is conducted through partnerships with established pharmaceutical companies, Anthropic's programme is run directly by the company with its own target selection and compound evaluation pipeline.
How can researchers apply for the Claude Science AI for Science grant programme?+
Anthropic is funding up to 50 Claude Science projects with grants of up to $30,000 in API credits each. The programme targets graduate and postdoctoral researchers rather than corporate teams. Applications are open through 15 July 2026, with award notifications sent by 31 July 2026. Funded projects run from 1 September to 1 December 2026. Indian researchers at IITs, IISERs, IISc, and other qualifying research institutions are eligible to apply through Anthropic's AI for Science programme page.
Written by
TechPillow Team
Sharing insights on technology, product development, and the Indian tech ecosystem.