OpenAI releases GPT-6-Cyber for approved security teams (November 5, 2026)

What you get

gpt-6-cyber is a new OpenAI model for security work. OpenAI says to use it to investigate a vulnerability, reproduce it in an authorized test environment, and verify a fix. It works on the Responses and Chat Completions APIs. The context window is 1,050,000 tokens, with up to 128,000 output tokens.

Price per 1M tokens: $20 input, $2 cached input, $100 output. Cache writes cost $25 for 30 minutes or $50 for 12 hours.

On November 5, 2026 OpenAI also deprecated gpt-5.5-cyber-preview. It will be removed from the API on February 1, 2027; OpenAI says to move to the most capable cyber model available to you.

It is not open to everyone. Your organization needs Daybreak Red access and approval.

Prerequisites

• Daybreak Red access and GPT-6-Cyber approval for your organization. Your OpenAI account team confirms this.
• An organization owner who enables Daybreak Red for the project you will use. Use a non-default project and a key from that project.
• An IP allowlist set up for the public IPs your app sends requests from.
• If the project uses Zero Data Retention, Private Safety Processing set up for it.
• An OpenAI API key from the approved project in OPENAI_API_KEY.

Steps

1. Confirm access and approval with your account team.
2. Ask an organization owner to enable Daybreak Red for the project.
3. Set up the IP allowlist for your app's public IPs.
4. If the project uses ZDR, set up Private Safety Processing. Storage should show "Validated".
5. Set OPENAI_API_KEY to a key from the approved project.
6. Send a request. This is the official example:

curl https://api.openai.com/v1/responses \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-6-cyber",
    "access_programs": {"cyber": "daybreak_red"},
    "safety_identifier": "user_a17f",
    "input": "Suggest a regression test for a path traversal fix."
  }'