On Oct 21, PHP published a security advisory fixing two critical unauthenticated remote code execution flaws (CVE-2026-109618, CVE-2026-109619) affecting every PHP install serving web requests, with mod_php at highest risk. OpenAI disclosed the same day that Calif found a flaw using GPT-5.6 Sol under its Patch the Planet initiative, with Wiz independently finding and reporting both; OpenAI coordinated a two-month disclosure and committed $1B in subsidized Daybreak access for under-resourced defenders.

Key Takeaways

  • ✓Flaws: CVE-2026-109618 (memory corruption via OOM handling) and CVE-2026-109619 (stale global state across requests), both unauthenticated RCE (PHP official)
  • ✓Fix: upgrade to PHP 8.5.12, 8.4.27, 8.3.36 or 8.2.35; PHP says there are no complete mitigations and updating is the only safe choice
  • ✓Scope: PHP-FPM, Apache mod_php, php-cgi, FrankenPHP and LiteSpeed are affected, with a demonstrated exploit on mod_php
  • ✓AI role: Calif found a flaw with GPT-5.6 Sol, Wiz reported both independently; OpenAI coordinated a ~2-month disclosure (vendor account)
  • ✓Investment: OpenAI committed $1B in subsidized Daybreak access and will ship a Codex skill to check PHP versions
OpenAI Patch the Planet: GPT-5.6 Sol helped find two critical PHP RCE flaws, fixed in PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35
🖼️Official Media
Click to view high-res
🧭

Turn your technical choice into a development budget

Compare 40 dev plans & simulate token costs vs $20/mo subscriptions

🔬

In-Depth Technical Analysis

Background and the pain point

PHP still powers a large share of websites and web apps; OpenAI's post puts it at roughly 70% of sites. Once an infrastructure-level flaw goes public, attackers often move faster than operators. As frontier models get better at finding bugs, the question shifts from whether a flaw can be found to whether fixes reach distributions, hosts and operators before attackers do.

Architecture and mechanism

Per PHP's advisory, CVE-2026-109618 lets out-of-memory handling interrupt PHP while request data is only partially initialized, and request cleanup then treats the invalid data as valid. CVE-2026-109619 lets request-owned internal state survive after its memory is freed, so a later request on the same worker may use it before initialization. Both allow unauthenticated remote code execution. Per OpenAI's blog, Calif found a flaw using GPT-5.6 Sol under Patch the Planet, and Wiz independently found and reported both. PHP maintainers led the fix, while OpenAI, Wiz, Cloudflare and others coordinated for about two months, including early protections with distributors, security vendors and US agencies such as ONCD and CISA.

Benchmarks

This is a security event, not a model launch: no model benchmarks or discovery-rate figures were published. The verifiable facts are that PHP-FPM, Apache mod_php, php-cgi, FrankenPHP and LiteSpeed are affected, and mod_php has a demonstrated exploit where an unauthenticated request triggers the attack before any application code runs. OpenAI calls it one of the most impactful security issues in recent memory; that is the vendor's characterization.

Getting started

Upgrade now to PHP 8.5.12, 8.4.27, 8.3.36 or 8.2.35; distributions and LTS vendors received patches in advance, so check for updated packages. PHP 8.1 and older are unsupported, but the security team published minimal source backports for 7.4, 8.0 and 8.1. If you cannot upgrade yet, PHP suggests memory_limit=-1 via php_admin_value, a low post_max_size and request-size limits at the web server or proxy, none of which makes an install safe. OpenAI is also sharing a PHP version checker and will ship a matching Codex skill.

Action HubReady to adopt this in production?

Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.

Up NextClampdown open-sources a hardened container sandbox for Claude Code, Codex, and more
Read Now→