They Changed the Contract
June 2, 2026 · 5 min read
You signed a one-year contract. You paid for 12 months upfront. You thought the price was locked.
On June 1, GitHub changed the terms — not by raising your subscription fee, but by changing what your subscription buys.
🔴 The Multiplier Stealth
What changed on June 1 for annual subscribers:
- Claude Opus 4.7 multiplier: 7.5x → 27x (3.6x increase)
- GPT-5.4 multiplier: 1x → 6x (6x increase)
- Legacy request quotas now draining up to 4x faster
- Your contract price didn't change — but what it buys did
Source: Notebookcheck (Darryl Linington, June 2, 2026)
How the Stealth Increase Works
GitHub Copilot uses a "multiplier" system: each AI model costs a certain number of "requests" per interaction. If Claude Opus has a 7.5x multiplier, one query uses 7.5 requests from your monthly allowance.
When they moved the multiplier from 7.5x to 27x, your monthly allowance stayed the same — but it now buys 3.6x less. Same contract. Same price. 360% more expensive for the exact same usage.
This isn't a price increase you'd notice in your billing dashboard. The dollar amount is the same. What changed is the invisible exchange rate between your money and the product.
📄 The Annual Subscriber Trap
Annual subscribers specifically chose to pay upfront for a year because they wanted predictable pricing. They locked in their rate. GitHub's response: change the multiplier so the "locked" rate buys 4x less.
And the refund window? It closed on May 20 — 12 days before the multiplier change took effect.
Source: Notebookcheck, Windows Forum (June 2, 2026)
The Enterprise Dashboard Doesn't Lie
A viral screenshot from an enterprise dashboard tells the story in one image:
💰 $500 → $5,291 in One Month
A team whose historical monthly usage billed at $500.35 recalculated to $5,290.92 under the new token metering. That's a 10.6x increase for the same team, same codebase, same usage.
The only thing that changed: the meter.
Source: Viral enterprise dashboard screenshot / Notebookcheck (June 2, 2026)
The Hidden Caps You Can't See
Even the "pay-as-you-go" framing is a trap.
🔒 You Can't Buy More Credits
GitHub implemented temporary spending caps tied to account age and verification status. Developers who hit their ceiling cannot purchase additional credits. The only path to restoring premium features? Upgrading to a higher subscription tier.
That's not usage-based pricing. That's an upsell funnel with extra steps.
Source: Notebookcheck (June 2, 2026)
The Pattern: Same Week, Two Vendors, Same Playbook
GitHub's multiplier stealth didn't happen in a vacuum. In the same week:
- June 1: Copilot switches to token billing — with hidden multiplier increases on annual subscribers
- June 1: Anthropic files for IPO at $965 billion valuation
- June 15: Anthropic splits agent SDK access into a separate credit pool — no rollover, no fallback, opt-in required
- Both vendors use the same word: "sustainable"
Two vendors. Same week. Same extraction pattern. Coincidence?
📈 The Subsidy Confession
Internal Microsoft documents (obtained by journalist Ed Zitron) show Copilot's running costs nearly doubled week-on-week since January 2026. The subsidy was always temporary. You were never the customer — you were the onboarding period.
Source: Ed Zitron / Microsoft internal documents, cited in Notebookcheck (June 2, 2026)
What This Means for You
If you're an annual Copilot subscriber, your contract now buys 4x less than it did last week. You can't get a refund. You can't buy more credits without upgrading. And the promotional buffers ($30/$70 extra through August) disappear in three months.
If you're on any AI subscription that uses tokens, credits, or multipliers: watch the multipliers, not the price. The vendor doesn't need to raise your rate to charge you more. They just need to change what your rate buys.
The Alternative: A Price That Doesn't Change
The Twin Agent Kit is $47 once. No credits. No multipliers. No meter. No annual plan that can be rewritten mid-contract.
You own it. The price doesn't change because there is no meter to change.
Stop renting AI. Start owning it.