OpenAILanguage modelLimited access
GPT-6 Cyber
Summary
GPT-6 Cyber is an OpenAI reasoning model built for authorized security research. It is meant to find and validate exploitable vulnerabilities within the scope that researchers set. It launched on November 5, 2026, at first only through the API and Codex for a selected group of approved Daybreak Red customers.
The model accepts text and images and answers in text; it supports neither audio nor video. Its context window holds 1,050,000 tokens and output is capped at 128,000. At short context, input costs $20 per million tokens, output $100, and cached input $2. Writing to the cache costs $25 for 30 minutes or $50 for 12 hours.
All specifications
Input and output
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
- Access
- Initially available through the API and Codex to a selected group of approved Daybreak Red customers.
- Source: Access
API
Further details
Prices
- Short-context input price
- $20 per 1M tokens. OpenAI labels this the standard short-context price.Source
- Short-context cached input price
- $2 per 1M tokens. OpenAI labels this the standard short-context price.Source
- Short-context output price
- $100 per 1M tokens. OpenAI labels this the standard short-context price.Source
- Cache write price, 30 minutes
- $25 per 1M tokens for 30 minutes. OpenAI specifies this for short contexts.Source
- Cache write price, 12 hours
- $50 per 1M tokens for 12 hours. OpenAI specifies this for short contexts.Source
Frequently asked questions about GPT-6 Cyber
How large is the context window of GPT-6 Cyber?
1.05 M tokens. That is how much text GPT-6 Cyber reads at once, about 1,580 A4 pages.
Page count as a rule of thumb from Artificial Analysis