Tools Model Security Capability gpt-daybreak-red
Model record
GPT Daybreak Red: how it does on security work, and what it will answer
- Lab
- OpenAI
- Released
- 2026-08-10
- Access
- gated
- Weights
- closed
- Safeguards
- Trusted Access for Cyber, offensive tier: separate approval, stronger monitoring, hardware security keys from 1 September 2026
- Score rows
- 0
- Confidence
- SINGLE
What happened
The only model in the field sold for "proof-of-concept exploit development, exploit-chain validation, penetration testing, and red teaming", at $12.50 and $75 per million tokens with a 400,000-token window. No public benchmark row exists for it.
Refusal gate (SINGLE, observed 2026-09-11): Refusals lifted for approved organisations under separate provisioning.
Sources
- OpenAI python client, model union naming gpt-daybreak-red-latestraw.githubusercontent.com/openai/openai-python/main/src/open…
- Refusal policy or licenceraw.githubusercontent.com/openai/codex/main/codex-rs/tui/src…
One record from the Model Security Capability, maintained by the Security Desk. Data: CSV · JSON ·RSS · CC BY 4.0 with attribution to YFarmX.Tracker updated · 19 September 2026