OpenAI Daybreak, run by LESIS
Your path to Daybreak.
OpenAI now puts a significant share of its frontier compute into finding and patching vulnerabilities, and keeps the models that do it for approved defenders. Daybreak is how those defenders get them.
LESIS is verified by OpenAI for Daybreak. We point those models at your code and systems, prove what they find and hand your team the fixes.
On the record
- 948potential issues in the Linux kernel, each with a patch already prepared.
- 81new vulnerabilities found in Firefox.
- 2 tiersDaybreak Blue for discovery and response; Daybreak Red for models trained for cybersecurity and exploit validation.
- $1 billioncommitted by OpenAI to the defenders of critical infrastructure.
Why it matters
Frontier compute is now pointed at vulnerabilities. The only question is who points it at yours first.
Daybreak gives approved defenders OpenAI's most capable models with safeguards tuned for security work. Daybreak Blue is for discovery and incident response. Daybreak Red adds models trained for cybersecurity, built for vulnerability research and exploit validation. Pointed at the Linux kernel, they prepared patches for 948 potential issues. Pointed at Firefox, they found 81 new vulnerabilities.
OpenAI's stated aim is to put frontier intelligence in the hands of trusted defenders before attackers can deploy offensive AI at scale, and it calls the time left to do that a narrowing window. The bugs these models find are already in your code. Someone will read it with a model like this. It should be you.
How LESIS runs it
You get the findings.
We carry the model.
Access, scope, compute and the engineers who check every result come from LESIS. What reaches your team is a short list of proven vulnerabilities, each with a way to fix it.
Scope and authorization
We agree on what is in scope, from repositories to applications and infrastructure, with the written authorization these programs require. Nothing runs outside it.
Found by AI, proven by LESIS
The model reads your code and systems the way an attacker would. LESIS engineers reproduce every result before it counts, so you get vulnerabilities, not suspicions.
Into your record
Every proven finding lands in the LESIS platform with its evidence, severity, an owner and a deadline, next to the rest of your program.
Fixed, then retested
When your team ships the fix, we run the check again and close the finding only when the hole is gone.
Access
Built for approved defenders.
Daybreak is open only to people and organizations OpenAI approves for authorized security work, and individual accounts must sign in with hardware security keys. LESIS has been through that verification and meets those requirements, and runs the work for you, so your team gets the results without the application or the overhead.
Sources
- OpenAI · Expanding Daybreak as the cyber defense window narrows
- CNBC · OpenAI expands Daybreak cybersecurity initiative
- TechCrunch · OpenAI released a new cybersecurity model
Claude Mythos is developed by Anthropic and Daybreak is a program of OpenAI. Their names are used here only to describe the access LESIS provides. Figures are as published by Anthropic and OpenAI or reported by the press.
Frontier models are already reading code. Make sure they read yours first.
Tell us what you run. We will tell you where Daybreak should look first and how soon it can start.
