www.it-connect.fr
OpenAI Launches Daybreak, a Cybersecurity Platform to Challenge Anthropic
OpenAI has just launched Daybreak, a cybersecurity platform built on its GPT-5.5 models and its Codex Security agent. The goal: to compete with Anthropic in vulnerability hunting.
Daybreak: AI Serving Security by Design
In recent weeks, we have heard a lot about Anthropic, especially Claude Mythos Preview. It must be said that it appears particularly effective at vulnerability research. The work carried out in partnership with Mozilla to uncover flaws in Firefox thanks to AI is a good example.
With Daybreak, OpenAI has chosen to adopt the following philosophy: security should not be a simple layer added after the fact, it must be built in from the very first lines of code. This initiative therefore aims to speed up code security analysis, generate and test fixes directly within a sandbox, and then provide companies with ready-to-use documentation.
From now on, OpenAI will offer its GPT-5.5 model in three tiers in order to be more flexible when it is used in a "cyber" context:
- GPT-5.5 : intended for general use.
- GPT-5.5 with Trusted Access for Cyber : designed for verified defensive workflows such as secure code review, malware analysis, vulnerability triage, and fix validation.
- GPT-5.5-Cyber : built for specialized workflows such as red teaming and penetration testing.
GPT-5.5 can also call on the Codex Security agent to perform cybersecurity-related analyses. According to OpenAI, this agent models threats using realistic attack paths and tests potential flaws before recommending remediation measures. It has already helped fix more than 3,000 vulnerabilities.
"AI now enables defenders to analyze entire codebases, identify the most subtle vulnerabilities, validate fixes, analyze unfamiliar systems, and move faster from detection to remediation. Because those same capabilities can be misused, Daybreak combines broad defensive capabilities with trust, verification, proportionate safeguards, and accountability. The goal is simple: help cyber defenders move faster and keep software secure at all times.", OpenAI explains.
Adoption by Several Major Companies
Although access to Daybreak remains limited and is available by request only, the platform is already reportedly attracting big names in IT and network security. OpenAI has named several companies such as Cisco, Cloudflare, CrowdStrike, Oracle, Akamai, and Fortinet (so, fewer vulnerable Fortinet products soon, do you think?).
On this page on OpenAI's website, you can fill out a form to request a vulnerability scan using Daybreak. It remains to be seen whether this type of initiative will allow OpenAI to generate substantial revenue. That seems necessary, because according to The Wall Street Journal (relayed by Le Figaro), OpenAI's financial situation may not be in great shape.
