OpenAI-led coalition coordinates global response to critical PHP vulnerabilities found by GPT-5.6 Sol

OpenAI led a coalition to address critical vulnerabilities in PHP, identified by GPT-5.6 Sol and independently by Wiz, which the company considers one of the most impactful security issues in recent memory.
A proposed five-stage framework coordinates the response, involving upstream validation, advance notice to affected organizations, government engagement, product updates, and public guidance to ensure defenders have time to prepare.
OpenAI committed $1 billion in subsidized Daybreak access for frontline defenders and provides Codex for OSS subscriptions to support maintainers in testing fixes and handling security reports.
The initiative aims to build secure-by-default software through AI-assisted coding, memory-safe migrations, formal verification, and stronger isolation to prevent future vulnerabilities.

Response Coordination

The framework engages U.S. government partners including the White House’s Super Intelligence Force, ONCD, and CISA to help guidance reach critical infrastructure.
Participants receive necessary information, from network logs to proof-of-concept exploits, with maintainers deciding whom to involve based on protective role and sensitivity.
The Patch the Planet website aggregates public disclosures, official patches, and interim mitigations to help operators verify updated software.

OpenAI provides funding, model access, and engineering help to small maintenance teams to separate new issues from duplicates and test fixes across releases.
Investments include direct sponsorship of projects like the Rust Foundation and reusable security workflows to reduce the effort required for future findings.
Additional blocking classifiers are being calibrated for Daybreak to balance broad access for defenders against the risk of misuse.