Contains affiliate links. We may earn a commission — it never changes the verdict.Details
The Stack / Coding
GitHub Copilot
03 CreationYes at $10 if you code in VS Code or JetBrains and want the lowest-friction AI assist — completions stay unlimited on every paid plan even after the June 2026 move to usage billing. For agent-heavy work, compare credit math against Cursor before committing.
The job in the stack
The completion seat: always-on code suggestions in your existing editor, plus agent work metered by AI credits.
Pricing
| Plan | Price | Notes |
|---|---|---|
| Pro | $10/mo | Includes $15 in AI credits; completions unlimited |
| Pro+ | $39/mo | Includes $70 in credits |
| Max | $100/mo | Includes $200 in credits |
| Business | $19/user/mo | $19/user credits included |
| Enterprise | $39/user/mo | $39/user credits included |
Pricing verified 2026-08-30. Plans change — the source of truth is GitHub Copilot's own pricing page.
Models & versions
- Claude Sonnet 5Introductory token rate $2/M in, $10/M out through Aug 31, 2026; reverts to the standard $3/M in, $15/M out on Sept 1 — a 50% jump in the per-token cost of agent work that draws on Sonnet 5, though completions stay unlimited regardless
- GPT-5.6 Sol / Terra / LunaIn the model picker since July 9, 2026. Terra and Luna reach Pro and up; Sol is gated to Pro+, Max, Business, and Enterprise. Business and Enterprise admins must switch the policy on — it ships off
- GPT-5.5Still selectable, now a tier below GPT-5.6
The honest read
What works
- +Unlimited completions on all paid plans — the core value is uncapped
- +Native to GitHub: PR summaries, code review, coding agent in issues
- +Cheapest serious entry point in AI coding at $10
What breaks
- −June 2026 usage-billing switch makes agent costs variable — 1 credit = $0.01
- −Editor experience trails Cursor for deep agentic sessions
- −Credit pricing differs per model — you need to actually read the rate card
- −Claude Sonnet 5's introductory token rate ends Sept 1, 2026 — a 50% jump ($2/$10 to $3/$15 per 1M tokens) for anyone running agent work on it
Use it for
- →Developers in stock VS Code or JetBrains who want zero setup change
- →GitHub-centric teams using PR and issue automation
- →Budget-conscious coders — $10 with credits included
Skip it if
- ×You want the deepest in-editor agent workflow — that's Cursor
- ×You don't use GitHub — half the integration value evaporates
Questions operators ask
?Did Copilot get more expensive with usage billing?
For completion-led users, no — completions stay unlimited and $10 Pro includes $15 of credits. For agent-heavy users, costs became variable: heavy months now bill like a utility.
?Can I use Claude models in Copilot?
Yes — Claude Sonnet 5 is available in Copilot. Through August 31, 2026 it runs at an introductory $2/M input, $10/M output; from September 1 that reverts to the standard $3/M in, $15/M out, a 50% increase per token. It only bites agent-heavy usage — completions stay unlimited on every paid plan either way.
What changed
- Closed a standing item: confirmed Claude Sonnet 5's introductory Copilot rate ($2/M in, $10/M out) expires August 31, 2026 and reverts to the standard $3/M in, $15/M out on September 1 — a 50% per-token increase for agent work drawing on that model. Completions remain unlimited on every paid plan regardless; plan prices and credit allotments unchanged.
- Re-verified — plan prices and included-credit amounts unchanged since the June 1 move to AI Credits. Model lineup updated: GPT-5.6 Sol/Terra/Luna entered the picker July 9, 2026, with Sol restricted to Pro+, Max, Business, and Enterprise. Completions confirmed still unlimited and credit-free on every paid plan.
- Re-verified — no changes since July 3. Pricing, credit allotments, and model lineup hold.
- Verified post-transition pricing: Pro $10 (+$15 credits), Pro+ $39 (+$70), Max $100 (+$200), Business $19/user, Enterprise $39/user.
- All plans moved to usage-based billing with monthly AI credit allotments. Completions remain unlimited on paid plans.
Give GitHub Copilot a job
The completion seat: always-on code suggestions in your existing editor, plus agent work metered by AI credits.
Try GitHub Copilot →Alternatives with different jobs
Cursor
03 CreationThe hands-on coding seat: an editor where AI agents write, refactor, and navigate the codebase with you.
Claude
02 DecisionsThe thinking seat: long documents, hard reasoning, code, and anything where the words have to be right the first time.