Anthropic released Claude Mythos Preview on April 7, 2026, its most capable model yet for coding and agentic tasks, priced at $25/$125 per million input/output tokens via the Claude API and major cloud providers. In security red-team testing, Mythos Preview autonomously developed working browser exploits 181 times—compared to 2 times for Opus 4.6—and identified thousands of zero-day vulnerabilities across operating systems, cryptographic libraries, and web infrastructure. Alongside the model launch, Project Glasswing was announced: a consortium of 50+ organizations including AWS, Apple, Broadcom, Cisco, Google, Microsoft, and NVIDIA, backed by Anthropic committing up to $100M in model usage credits and $4M in donations to open-source security organizations. Glasswing aims to use Mythos to systematically audit critical infrastructure before adversaries can exploit the same capabilities.