Anthropic Grants Indian Entities Access to Mythos
Anthropic expanded its Project Glasswing program on June 2, 2026, granting access to Claude Mythos Preview to roughly 150 additional organizations across more than 15 countries, including India, according to the company's blog post. An earlier cohort of about 50 partners has already used Mythos to scan codebases and found more than 10,000 high- or critical-severity security flaws, Anthropic said. The expansion targets organizations maintaining critical infrastructure in power, water, healthcare, communications, and hardware, where Anthropic estimates a successful attack could affect more than 100 million people per partner. TechCrunch and NDTV covered the expansion; NDTV noted Anthropic has not disclosed which Indian organizations received access. The announcement came the same week Anthropic confidentially filed a draft S-1 with the SEC.
Anthropic's decision to more than triple the partner base for its specialized vulnerability-hunting model signals that AI-driven code auditing is moving from an experimental capability into a standard tool for defending infrastructure that, by Anthropic's own estimate, over 100 million people per partner depend on.
What happened
Anthropic announced in a June 2, 2026 blog post that it is expanding Project Glasswing, extending access to Claude Mythos Preview to approximately 150 additional organizations across more than 15 countries, including India. An earlier cohort of roughly 50 partners had already been using Mythos to scan codebases, and Anthropic said those partners had found more than 10,000 high- or critical-severity security flaws. Anthropic wrote, "What each partner has in common is that a successful attack on their codebase could be catastrophic," and estimated that for most partners a major attack could affect more than 100 million people. TechCrunch, citing the Financial Times, reported that the expanded group covers power, water, healthcare, communications, and hardware, and named several organizations given access in other countries; NDTV reported that Anthropic has not disclosed the names of the Indian organizations granted access.
Technical context
Anthropic describes Claude Mythos Preview as a specialized model for advanced cybersecurity research, code analysis, and vulnerability discovery, built around the Project Glasswing initiative for critical-infrastructure partners. TechCrunch noted that rival vendors are moving in the same direction: OpenAI's GPT-5.5-Cyber, a limited-preview model for vetted cybersecurity teams, has separately reported an 85.6% score on the CyberGym benchmark for reproducing known vulnerabilities, up from 81.8% for the base GPT-5.5 model. Security-focused models that automate static and dynamic code analysis can scale discovery of complex vulnerability classes, but they also change triage workloads and false-positive profiles for the practitioners who act on the findings.
For practitioners
Wider access to models like Claude Mythos Preview should accelerate discovery of supply-chain and maintainer-level vulnerabilities, raising the volume of high-priority findings that incident-response teams must handle. Teams evaluating these tools should compare detection coverage and false-positive rates, and plan for how automated findings integrate into existing CI/CD pipelines, bug-bounty programs, and coordinated-disclosure workflows rather than assuming outputs can go straight to remediation.
What to watch
Watch for which Indian organizations receive access and whether any named partners publish coordinated-disclosure reports; how partners operationalize Mythos outputs into triage, human review, and vulnerability-management pipelines; and whether other AI labs widen distribution of security-focused models in a way that pushes the industry toward standardized responsible-access controls. Separately, Anthropic's confidential draft S-1 filing with the SEC the same week - part of a broader wave of AI IPO activity alongside OpenAI and SpaceX - is worth tracking as a signal of the company's financial trajectory, though it does not by itself indicate an IPO timeline.
Key Points
- 1Anthropic expanded Project Glasswing access to about 150 organizations in over 15 countries, including India, widening its security model's user base.
- 2Partners using Claude Mythos Preview have found more than 10,000 high- or critical-severity security flaws in critical-infrastructure codebases.
- 3The expansion coincides with Anthropic's confidential draft S-1 filing with the SEC, signaling a broader push toward a public listing.
Scoring Rationale
A major AI lab meaningfully expanding a specialized vulnerability-discovery model to critical infrastructure across roughly 150 organizations in 15+ countries, with over 10,000 flaws already found, is a substantive security development for practitioners protecting essential services. It sits at the high end of the notable tier given its scale and Anthropic's concurrent confidential S-1 filing, without applying any recency discount.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems