Skip to content
Back
PRICING BRIEFINGDaily · Oct 17, 2026
GGitHub
GitHub
PACKAGING CHANGE

GitHub extends AI secret detection to push protection, both usage-based

GitHub renamed Secret Protection's "Copilot secret scanning" feature to "AI detection for alerts" — it still scans existing code the same way, just under a new name.

GitHub’s pricing page · Change Detected

WHY IT MATTERS

For teams already paying extra for Advanced Security as a standalone add-on since July 2025, this stacks another usage-based line item onto what used to be a single security line in the budget. Watch whether GitHub keeps extending usage-based billing into other security products next, the way it spread from Copilot into the cloud sandbox and now Secret Protection.

BY THE NUMBERS

  • GitHub's first pricing change in 131 days, vs. a ~105-day average gap
  • Third usage-based billing move since the May Copilot shift
  • Advanced Security became a paid add-on in July 2025
  • 9 of 53 tracked comparisons have been real changes since April 2024

THEIR RATIONALE

GitHub frames this as closing a detection gap: regex-based scanning catches structured tokens and keys, but plain-text passwords and other unstructured secrets slip through until a human catches them. AI detection for push protection stops them before the push completes rather than flagging them after the fact. This is GitHub's first pricing change in 131 days, well past its roughly 105-day average gap, and its third move tied to usage-based billing since it shifted Copilot itself onto that model earlier this year.

TRACK RECORD9 changes since Apr 1, 20244 in the last 12 monthsa change every ~4 monthsprevious change Jun 8, 2026 (131 days ago)

THE SIGNAL

The pattern matters more than the single feature: GitHub has now put three separate things on usage-based billing since the Copilot shift — the billing model itself, June's cloud sandbox add-on, and now AI secret detection. That's a deliberate push to meter AI-adjacent features by consumption rather than bundle them into a flat seat price. Across the wider index, 201 pricing changes at 153 companies have touched credits-and-usage billing in the last 180 days, about one in nine of all tracked changes, so GitHub metering a security feature this way fits a broader market drift away from flat pricing. The twist here is that this isn't optional tooling — it's a build-time security control, which raises the stakes of billing it by usage.

CREDITS & USAGE BILLING153 companies in the last 6 months11.6% of 1,726 tracked changes3× on GitHub's pageInvideo most often (4)

IN CONTEXT

GitHub has leaned into usage-based billing since shifting Copilot itself onto that model in May, following up a month later with a usage-priced cloud sandbox add-on for Copilot workflows — this is the third move in that pattern. It lands 131 days after GitHub's last pricing change, well past its roughly 105-day typical cadence, and continues a history of carving security features out of flat pricing: Advanced Security left the Enterprise plan for a paid add-on in July 2025. Across the wider index, 201 pricing changes at 153 companies have touched credits-and-usage billing in the last 180 days — about one in nine of all tracked changes — so metering an AI security feature by usage fits a market-wide drift away from flat seat pricing.

WHAT PEOPLE ARE SAYING

@LiuweijiaViptwitter

False positives are why people disable these checks. If the context part actually works, this one stays on.

View original
@Nivish_twitter

A security check can cost credits even when it blocks nothing. @github plans credit billing for opt-in AI push protection, while existing AI secret alerts stay included in Secret Protection. Check the feature, not just the label. https://github.blog/changelog/2026-10-07-purpose-built-model-for-leaked-secret-detection/

View original
@SupersocksInteltwitter

GitHub's new model for leaked secrets comes in two parts, and only one of them is free (changelog, Oct 7). Included in Secret Protection / Advanced Security: - AI-detected secret alerts, already switched to the new model Billed in AI Credits once you opt in, starting in the coming weeks: - AI checks in push protection, charged to the org that owns the repo - secret checks in Copilot's /security-review, charged to your Copilot plan The two details that change the bill: a push check can consume credits even when it doesn't block the push, and budget alerts alone don't stop usage. For a real cap, set a Secret Protection AI Credits budget with "Stop usage when budget limit is reached". If you're on the push protection private preview, continued use after billing starts consumes credits. Disable it beforehand if you don't want that.

View original