Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software.
— Anthropic (@AnthropicAI) April 7, 2026
It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans.https://t.co/NQ7IfEtYk7
They claim Claude Mythos Preview can find vulnerabilities better than all but the most skilled humans, which is a massive technical statement if it holds up in independent benchmarks.
A statement from Anthropic CEO, Dario Amodei, on our discussions with the Department of War.https://t.co/rM77LJejuk
— Anthropic (@AnthropicAI) February 26, 2026
On messaging, this launch is polished, high-stakes, and intentionally fear-adjacent: critical software, urgent initiative, frontier model.
A statement on the comments from Secretary of War Pete Hegseth. https://t.co/Gg7Zb09IMR
— Anthropic (@AnthropicAI) February 28, 2026
Now the knife fight: OpenAI and everyone else are also framing frontier models as practical infrastructure tools, not just chat toys.
Introducing GPT-5.5
— OpenAI (@OpenAI) April 23, 2026
A new class of intelligence for real work and powering agents, built to understand complex goals, use tools, check its work, and carry more tasks through to completion. It marks a new way of getting computer work done.
Now available in ChatGPT and Codex. pic.twitter.com/rPLTk99ZH5
My launch scorecard: Tech: 8.7/10 (big claim, plausible, but needs third-party validation), Comms: 9.1/10 (crisp story, high urgency, strong framing), Pricing: 5.8/10 (no clear public number in the launch post, and uncertainty always drags score), Hype-vs-Substance: 7.4/10 (more substance than most AI launches, still leaning on broad superiority language), Competitive Position: 8.3/10 (excellent enterprise angle, but crowded frontier field with aggressive rivals).
Bottom line: Glasswing is not fluff, but it’s also not a checkmate yet. If Anthropic ships transparent benchmark tables, publishes real-world vuln detection lift (for example, +25% to +40% versus top human-team baselines), and avoids enterprise sticker shock, this could become the first “frontier model” story that security leaders treat like infrastructure instead of marketing.
Stay sharp. — Max Signal