the diff / OpenAI API / November 5, 2026

OpenAI ships GPT-6-Cyber. Daybreak Red approval required

score approvals before the first request: "2"

OpenAI's November 5 changelog adds gpt-6-cyber, $20 in and $100 out per Mtok, 1,050,000 context, for orgs with Daybreak Red plus model approval.

What changed

OpenAI's API changelog for November 5 lists GPT-6-Cyber, model id gpt-6-cyber, "to investigate vulnerabilities, reproduce them in an authorized test environment, and verify fixes." The model page prices it at $20 per million input tokens, $2 cached and $100 output: 1,050,000 token context, 128,000 max output, Chat Completions and Responses only. The Daybreak guide lists it as "Red + additional model approval": Daybreak Red access plus a separate approval for this model.

Why it matters if you ship with an agent

Most organizations cannot call it, so do not put the id in a config and hope. If you are approved, OpenAI's setup guide asks for a non-default project with Daybreak Red enabled, an IP allowlist for the addresses your app sends from, Private Safety Processing on any Zero Data Retention project, and a stable pseudonymous safety_identifier per user. Budget for output at five times the input rate.

By MK · Source: https://developers.openai.com/api/docs/changelog

more OpenAI API changes: OpenAI API adds 12-hour prompt caching at 2.5x write price · OpenAI API usage tiers go from five to three · all

keep handy: AI coding tool pricing tracker · vibe coding vs AI coding · agent fail hall of fame


The Diff is Accept All's short note on changes that alter how you ship with an agent. Newer: Claude Code 2.1.319 can deny instead of prompting, auto mode loses fallback. Older: Claude text watermark now covers Mythos 5, Opus 4.7 and Sonnet 4.6.