India is preparing to test Anthropic’s cybersecurity-focused AI model Claude Mythos, with major Indian technology companies including Infosys and Tata Consultancy Services expected to receive access to Mythos Preview, according to a Bloomberg report.
The Indian government has established a secure testing facility and identified specific sectors and experts to evaluate the model’s performance and associated risks. This initiative comes as Anthropic has restricted public release of Claude Mythos due to significant cybersecurity concerns, limiting access through Project Glasswing to select organizations such as Apple, Nvidia, AWS, and CrowdStrike.
Infosys will specifically test Mythos’s capabilities for creating security fixes for its widely used banking platform Finacle. Tata Consultancy Services will similarly test its software systems for potential threats and vulnerabilities. Neither company currently has access to Mythos and is using Claude Opus 4.7 to patch vulnerabilities in the meantime, according to Business Today.
India is actively negotiating access to Claude Mythos in a secure setup, preferably within the country, with US officials led by the Ministry of External Affairs. The negotiations reflect growing concerns about the model’s capabilities after it demonstrated expert-level cybersecurity performance, outperforming humans at hacking tasks and identifying vulnerabilities including one that had persisted for 27 years with minimal oversight.
The Indian government has already asked banks to stay on high alert over Claude Mythos due to concerns about its potential to exploit vulnerable systems. This precautionary measure comes after the model was announced in early April 2026 and quickly gained attention for its advanced cybersecurity capabilities.
India was not among the initial 12 Project Glasswing launch partners, meaning no Indian bank, government agency, or tech company had access to the model until these recent negotiations began. The country’s push for access underscores the growing importance of AI-powered cybersecurity tools in protecting critical infrastructure and financial systems.
The testing initiative represents a significant step in India’s approach to emerging AI technologies, balancing the potential benefits of advanced cybersecurity capabilities with the need to manage associated risks. As the negotiations continue, the outcome will likely influence how other countries approach access to similarly powerful AI models with dual-use potential.
Read Article: TCS launches SovereignSecure Cloud in Europe
