GPT-5.6 Cyber

Closed weights OpenAI August 2026

No estimate

No hardware requirements for this model

The weights for this model have not been published, so it cannot be downloaded or run on your own hardware at any size. It is reachable only through its provider, and no graphics card changes that.

On record

Full specification

Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.

Origin

Who built this model, where, and when it was published.

Organisation
OpenAI
Organisation type
Industry
Country
United States of America
Published
11 August 2026

What it does

The problem areas the model was built for. A model can carry several of each.

Domain
Cybersecurity, Language
Task
Coding, Question answering, Software vulnerability discovery and exploitation
Base model
GPT-5.6 Sol

Size

How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.

Training data
tokens

Availability

Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.

Weights
Closed — provider access only
Model access
API access
Training code
Unreleased

Available through OpenAI's "Daybreak Red" program, the more restricted of two tiers of their cybersecurity-specialized services which launched alongside GPT-5.6 Cyber.

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Why it is tracked
Discretionary

Based on 5.6 Sol and thus assumed to be near-frontier general capability at time of release with specialized cyber capabilities.

Record confidence
Confident

Sources

Where this record came from and when it was last checked.

Reference
Expanding Daybreak as the Cyber Defense Window Narrows
Last updated
11 August 2026

What the numbers mean

Background

GPT-5.6 Cyber was published by OpenAI, in the country recorded as United States of America, during August 2026. The publishing organisation is categorised as industry.

It works in the domain of Cybersecurity, Language, and is recorded as performing the task of coding, Question answering, Software vulnerability discovery and exploitation.

Its starting point was an existing base model, GPT-5.6 Sol. That is why it shares the base model's general shape and size.

Its weights were never published, so it can only be reached through its provider. No graphics card changes that.

What went into building it

It is tracked in the underlying dataset for one reason in particular: discretionary.

Answers

GPT-5.6 Cyber — common questions

01

GPT-5.6 Cyber— when was it released?

It was published in August 2026.

02

GPT-5.6 Cyber— what is it used for?

It works in the domain of Cybersecurity, Language, and is recorded as handling the task of coding, Question answering, Software vulnerability discovery and exploitation. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.

03

GPT-5.6 Cyber— what GPU do I need to run it?

None. This is a closed model — its weights were never published, so it cannot be downloaded or run on your own hardware at any price. It is reachable only through its provider.

04

GPT-5.6 Cyber— is it open source?

No. Its weights have not been published, so it exists only as a service controlled by its owner.

05

GPT-5.6 Cyber— how many parameters does it have?

No parameter count has been published for it, which is why no memory or speed figure appears on this page.

06

GPT-5.6 Cyber— who created it?

It was published by OpenAI, based in United States of America, an organisation categorised as industry.

Source

Original publication

Record last updated 11 August 2026

The other direction

Looking at it from the other side?

This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.