Claude Opus 5 in GitHub Copilot is metered, not included
GitHub added Claude Opus 5 to Copilot on July 24, billed at provider API list price under usage-based billing. Your Copilot bill now has two halves.
Claude Opus 5 is now available in GitHub Copilot, and the interesting line in the changelog isn't about benchmarks. It's the billing sentence: Opus 5 is billed at provider API list price under usage-based billing. Your Copilot seat does not cover it. If your team turns it on and nobody says anything, you find out at the end of the month.
What actually happened
GitHub announced on July 24, 2026 that Opus 5 is available to Copilot Pro+, Max, Business, and Enterprise users, rolling out gradually across VS Code, Visual Studio, Copilot CLI, github.com, JetBrains, Xcode, Eclipse, and mobile. Anthropic shipped Opus 5 the same day at $5 per million input tokens and $25 per million output — half of Fable 5, and the same rate as Opus 4.8.
Two details matter more than the model card:
- No multiplier. GitHub isn't reselling Opus 5 at a premium request rate. It's pass-through at Anthropic's list price, which means your Copilot spend is now partly a function of how chatty your agents are, not how many people you employ.
- Cyber guardrails ship with it. GitHub notes enhanced safeguards for high-harm cyber content that may block certain security-related requests. If your team uses Copilot for vulnerability triage or exploit analysis, expect refusals.
Why metered AI coding matters for your business
Seat pricing is predictable and you can forecast it. Token pricing is a variable cost attached to a tool that developers are actively encouraged to run in long autonomous loops. Those two facts point in opposite directions.
The failure mode isn't a runaway bill in month one — it's that you lose the ability to say what an engineering hour costs. A five-person team on Copilot Business used to be a fixed line item. Now it's a fixed line item plus a metered one that scales with agent runs, and the two aren't attributable to the same thing.
So instrument it before you adopt it. Set org-level budgets in Copilot, decide which repos or tasks justify an Opus-class model versus a cheaper one, and put the number in front of whoever owns the budget monthly. Model choice is a dial you should be turning deliberately, the same way you'd size a database instance — not something each developer picks from a dropdown without seeing the meter.
Key takeaways
- Claude Opus 5 landed in GitHub Copilot on July 24, 2026 for Pro+, Max, Business, and Enterprise plans
- It's billed at Anthropic's API list price ($5/$25 per million tokens) under usage-based billing — your seat doesn't cover it
- No premium multiplier, but no ceiling either: cost scales with agent runs, not headcount
- Enhanced cyber-content safeguards may block legitimate security work — keep a second path for that
- Set budgets and pick models per task before you turn it on, not after the first invoice
Metered AI turns engineering into a variable cost. We help teams price what their automation actually consumes before they commit to it. Run the numbers on your stack.
Sources: GitHub Changelog, CNBC.
- #github-copilot
- #claude-opus-5
- #ai-coding-tools
- #dev-spend
- #usage-based-billing
Tommy Rush — Founder, Rush Commerce
Operator turned builder. 15+ years running operations — now shipping the systems businesses run on. More
Get The Rush Report weekly — one email, zero fluff.
Keep reading
Cursor's ₹649 India tier: dev seats go geographic
Cursor launched a ₹649/month India-only plan, roughly a third of its $20 Pro tier. AI dev tool pricing is now segmented by market, and that changes your budget math.
Read itvBulletin RCE exploit went public: get to 6.2.2 today
A working exploit for vBulletin CVE-2026-61511 dropped four weeks after the patch. Unauthenticated remote code execution, no in-the-wild reports yet. That window closes fast.
Read it