On June 1, 2026, GitHub Copilot pricing stopped being a flat subscription and became a token metering system. GitHub’s announcement put it simply: Copilot usage now consumes GitHub AI Credits, priced from each model’s published API rates. For the majority of developers — those using it primarily for autocomplete and occasional chat — the practical difference is close to nothing. Code completions stay unlimited. Next edit suggestions stay unlimited. The day-to-day experience does not change.
For a specific group — annual plan holders who run extended frontier-model sessions through Copilot Chat, or teams using Copilot’s agent mode heavily — the math is genuinely alarming. Annual subscribers stayed on premium requests, and GitHub raised their model multipliers on June 1. GitHub’s annual-plan multiplier table now charges 27 premium requests for one Claude Opus 4.8 request and 57 for one GPT-5.5 request.
Both of those statements are true at the same time.
How the new credit system actually works
Since June 1, every monthly GitHub Copilot plan runs on AI credits. One AI credit equals $0.01 USD. Every interaction that uses a premium model — a chat message, an agent session, a code review request — consumes credits based on input, output and cached tokens at the model’s API rate.
Two weeks before the switch, GitHub sweetened the individual plans. Its flex allotments announcement added a variable “flex” pool on top of base credits that match the price, plus a new Max tier:
| Plan | Price | Base credits | Flex allotment | Total monthly usage |
|---|---|---|---|---|
| Copilot Pro | $10/month | $10 | $5 | $15 |
| Copilot Pro+ | $39/month | $39 | $31 | $70 |
| Copilot Max | $100/month | $100 | $100 | $200 |
| Copilot Business | $19/user/month | $19 | — | $19, pooled |
| Copilot Enterprise | $39/user/month | $39 | — | $39, pooled |
(Source: GitHub Copilot plans and the GitHub Blog, verified October 2026)
Base credits are used first, then flex. GitHub says flex allotments “will vary over time,” so treat the base number as the only guaranteed one. Business and Enterprise credits pool across the organisation instead of sitting with each user.
GitHub launched a preview bill in early May so developers could model their consumption before the switch. The same Billing Overview page now shows your real credit usage, which is the number to check before changing plans.
What stays unlimited — and what does not
This distinction is where most of the noise about the pricing change gets muddled.
Unlimited, regardless of credits:
Code completions (the primary autocomplete feature) and next edit suggestions remain unlimited on all paid plans. If you are using Copilot mainly for inline code suggestions — the thing most developers spend most of their time on — your bill does not change.
Credit-consuming:
Copilot Chat interactions, agent mode sessions, code review requests, and multi-turn conversations all consume credits at per-token rates based on the model selected. The longer the session, the larger the context you are providing, and the more capable the model — the more credits it costs. If pull request review is most of your credit spend, the AI code review tools comparison covers CodeRabbit and Greptile as flat-priced alternatives.
This is the distinction that matters. A developer who uses autocomplete for eight hours a day and sends two short chat messages is paying roughly the same as before. A developer running three-hour Claude Opus 4.7 agent sessions through Copilot is about to see a very different bill.
The 27x multiplier explained — who it actually hits
The multiplier figure comes from annual plan mechanics. Annual Pro and Pro+ subscribers were not moved to credits. GitHub’s announcement says they “remain on their existing plan with premium request-based pricing until their plan expires,” and that their model multipliers increased on June 1. At expiry they drop to Free unless they pick a paid monthly plan.
A multiplier is how many premium requests one prompt burns. These are the current annual-plan values from GitHub Docs:
| Model | Premium requests per prompt |
|---|---|
| GPT-5.5 | 57 |
| Claude Opus 4.8 | 27 |
| Copilot code review | 13 |
| Claude Sonnet 4.6 | 9 |
| GPT-5.4 / Gemini 3 Pro | 6 |
| Claude Haiku 4.5 | 0.33 |
(Source: GitHub Docs, Model multipliers for annual plans, October 2026)
The people this hits hardest share a specific profile: they are on annual plans (so they committed to pricing based on old assumptions), they use Copilot Chat with frontier models regularly, and they run long sessions — code review of large PRs, agentic feature implementation, extended debugging. A 27x or 57x model drains a monthly premium request allowance in a handful of prompts. GitHub’s docs also say legacy annual plans won’t get access to new models and features.
If you are a monthly subscriber and you use Copilot mainly for autocomplete, the multiplier table has nothing to do with your bill.
GitHub Copilot is multi-model now — and the model choice is a cost decision
This is the aspect of the 2026 Copilot that deserves more developer attention: it is no longer a single-model product. Copilot Chat and agent mode let you pick between model families from Anthropic, OpenAI and Google — the annual-plan table alone lists Claude Haiku 4.5, Sonnet 4.6 and Opus 4.8, GPT-5.4 and GPT-5.5, and Gemini 3 Pro — and GitHub says only non-legacy plans get new models as they arrive.
Under usage-based billing, your model selection is a direct cost lever. Credits are charged at each model’s API rate, so the same task through a frontier model and through a small fast model can differ by an order of magnitude. If you are doing work where frontier coding performance matters — the kind of complex multi-file refactoring where architectural reasoning is the differentiator — the expensive model is the right call. If you are doing chat-based code explanation on a large file where any capable model is good enough, a cheaper model costs significantly less per session.
On legacy annual plans, GitHub’s docs also give a 10% multiplier discount when you let auto model selection choose. For routine chat that is a free saving.
Copilot vs Cursor vs Claude Code — the 2026 cost map
The pricing change shifts where Copilot sits in the competitive landscape in ways that are non-obvious.
Copilot Pro at $10/month is the clearest value in the market for developers who primarily use autocomplete. It works in every editor, completions are unlimited, and the $15 of credits covers occasional chat. If your main use case is inline suggestions and you do not run heavy agent sessions, this is still an excellent deal.
Copilot Pro+ at $39/month looked awkward at launch with $39 of credits. With the flex allotment it now carries $70 of usage, which makes it a reasonable middle tier for regular agent sessions. As covered in the AI tool comparison post, Cursor’s parallel agent mode is a tightly integrated product; Copilot’s agent mode is a layer on top of a UI built primarily for something else.
Cursor changed its own pricing in October. New customers now buy The X Subscription: Plus at $8 is Grok-only, and Premium at $30 bills Claude and GPT at API list price on top, according to Cursor’s pricing page. The Cursor X Subscription breakdown covers who should switch. If you want Claude models inside your editor, Copilot’s credits now cover them where Cursor’s new plans do not.
Claude Code remains separate from Copilot entirely. It runs in the terminal and operates autonomously, and Anthropic’s pricing page lists it as included in all paid Claude plans, starting with Pro at $20 a month. If you use Claude Code with an API key instead, the reducing Claude API costs guide covers how to keep that bill down.
The practical advice now that credits are live
Three things worth doing:
Check your real usage on the Billing Overview page. It shows credit consumption against your base and flex allotments. If you finish most months inside Pro’s $15, stay there. If you regularly buy extra credits, compare that spend to Pro+ at $70 of usage or Max at $200.
Check which model your Copilot Chat is using. Under per-token billing, leaving a frontier model selected for routine questions is the easiest way to burn credits you did not intend to spend.
If you are on an annual plan, look at the multiplier table before your next long session. Converting to monthly before expiry gives you prorated credits for the remaining value of the annual plan, according to GitHub’s announcement.
The headline “GitHub Copilot pricing unchanged” and the reality of “annual plan holders paying 27 or 57 premium requests per frontier-model prompt” are both technically true. The difference is whether you are using the product the way it was marketed or the way the most capable parts of it actually work at scale.
Pricing figures from the GitHub Copilot plans page, the GitHub Blog and GitHub Docs, re-verified October 2026.
(Updated October 15, 2026: corrected Copilot credit amounts after GitHub added flex allotments (Pro now includes $15 of usage, Pro+ $70) and the $100 Max plan; added Enterprise at $39 per user; replaced the third-party “27x cost multiplier” claim with GitHub’s official annual-plan multiplier table (27x applies to Claude Opus 4.8, 57x to GPT-5.5); corrected Claude Code billing, which is included in Claude’s paid plans rather than charged to API credits; removed unverifiable model token-efficiency and SWE-bench figures; and noted Cursor’s October 2026 move to The X Subscription.)