Anthropic brings Claude to AWS
- Anthropic announced Project Glasswing, giving AWS customers early access to Claude Mythos for defensive security tasks. - The project frames Claude Mythos as an AI-driven tool for vulnerability hunting and defensive security workflows. - Positioning AI as a defensive layer may change how security teams integrate model outputs into continuous detection and response (x.com).
Anthropic has put its new Claude Mythos model into Amazon Web Services, giving selected customers early access for defensive security work through Project Glasswing. (anthropic.com) Amazon Web Services said on April 7 that Claude Mythos Preview is available in Amazon Bedrock as a gated research preview, with access prioritized for defensive cybersecurity use cases. AWS described the model as able to analyze large codebases, identify sophisticated vulnerabilities, and show whether a flaw is exploitable. (aws.amazon.com) Anthropic said Project Glasswing launched with Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks. The company said it also extended access to more than 40 additional organizations that build or maintain critical software infrastructure. (anthropic.com) Security teams use vulnerability hunting to look for software bugs before attackers do, and continuous detection and response means folding those findings into day-to-day monitoring and patching. Anthropic said Glasswing is meant to put Mythos Preview into those defensive workflows so partners can scan first-party and open-source systems and fix weaknesses earlier. (anthropic.com) Anthropic said it is committing up to $100 million in Mythos Preview usage credits for the effort, plus $4 million in direct donations to open-source security organizations. The company said it does not plan to make Mythos Preview generally available at this stage. (anthropic.com) The caution reflects what Anthropic says the model can do. In a technical write-up, the company said Mythos Preview was capable in testing of identifying and exploiting zero-day vulnerabilities in every major operating system and major web browser, and said more than 99% of the vulnerabilities it found had not yet been patched. (red.anthropic.com) Anthropic’s system card says Claude Mythos Preview is its most capable frontier model to date and shows a large jump over Claude Opus 4.6 on several benchmarks. Three days after the Glasswing announcement, Anthropic released Claude Opus 4.7 and said it would test new cyber safeguards first on models that are less capable than Mythos Preview. (anthropic.com, anthropic.com) Amazon’s role matters because Bedrock is the company’s managed platform for building and deploying generative artificial intelligence applications, which gives Anthropic a controlled enterprise channel instead of a broad public release. AWS said the Bedrock listing for Mythos Preview is a research preview, not a general launch, and pointed users to a model card that says access is gated. (aws.amazon.com, docs.aws.amazon.com) The immediate test is whether security teams treat model output as a lead generator or as part of production defense operations. Anthropic said Glasswing is a starting point for sharing what partners learn, while keeping the model itself tightly limited. (anthropic.com)
