Anthropic has expanded Project Glasswing to grant Mythos access to 150 additional organizations across 15 countries, bringing total participating companies to about 200—including India, Samsung, Okta, Euroclear, and NYSE — adding critical infrastructure sectors like power, water, healthcare, and communications previously underrepresented.
The thinking behind this expansion is urgent defense: Mythos Preview has already found thousands of high-severity zero-day vulnerabilities, including a 27-year-old bug in OpenBSD and a 16-year-old flaw in FFmpeg, autonomously developing exploits that would take human professionals days to create.
Future expansion will fan out as more organizations meet security requirements, with NATO and ENISA joining, signaling government-level recognition that securing critical infrastructure is a top national security priority; Mythos Preview becomes available at $25/$125 per million tokens after initial credits expire.
Calibrated access is immensely critical to improving global cybersecurity because Mythos identifies dormant bugs in decades-old code and fixes vulnerabilities at impossible pace, giving defenders durable advantage as AI shrinks vulnerability-to-exploitation window from months to minutes.
AI DEFENSES CALIBRATED TO SHIELD WORLD FROM CYBERTHREATS.
Sanjay Sahay
Have a nice evening.