Is GitHub Copilot Pro worth it after metered billing? How far $15 of AI credits goes
Since June 1, 2026, GitHub Copilot bills chat and agent work in GitHub AI Credits instead of a flat request count. The change drew loud pushback from developers (see TechCrunch and The Register below). Copilot Pro still costs $10 a month. The question now is how much real work its credits cover.
What Pro includes (sourced)
- Price: $10/month.
- Credits: $15 of GitHub AI Credits a month: $10 base credits plus a $5 "flex allotment". GitHub says flex allotments "may change over time".
- Credit value: 1 AI credit = $0.01, so Pro's allowance is 1,500 credits a month. The allowance resets monthly.
- Free forever on paid plans: code completions and next edit suggestions don't use credits and stay unlimited.
- What spends credits: agent mode, code review, Copilot in the cloud, Copilot CLI and the Copilot app.
GitHub's FAQ says a quick question to a lightweight model costs "a fraction of a credit", while a long agent session on a frontier model across many files "costs more". It doesn't publish a fixed number per task. That's what we're here to measure.
When the credits run out
Per GitHub's plans page, on individual plans you can:
- Wait for the allowance to reset next cycle,
- Set a dollar budget for extra usage so Copilot keeps working. Credits draw down at $0.01 each, so a $10 budget covers 1,000 credits.
- Switch to a cheaper, lightweight model to stretch what's left.
You get usage alerts at 75%, 90% and 100% of any budget you set.
The plan ladder
| Plan | Price | Base credits | Flex | Total AI credits | Credits per $1 paid |
|---|---|---|---|---|---|
| Free | $0 | Limited chat & agent usage, 2,000 completions/month, entry-level models | — | ||
| Pro | $10/mo | $10 | $5 | $15 | 1.50× |
| Pro+ | $39/mo | $39 | $31 | $70 | 1.79× |
| Max | $100/mo | $100 | $100 | $200 | 2.00× |
The last column is our arithmetic: total credit value divided by price. Flex amounts can change, so this ratio can too.
Hands-on: how far $15 goes
Test brief: [BRIEF v1 — browser game: title screen, score, restart]
- Credits used by one full agent-mode build on the default model: [TEST RESULT PENDING]
- Same build on a lightweight model: [TEST RESULT PENDING]
- Builds that fit in one month's $15: [TEST RESULT PENDING]
- Credits used by one code review on a ~300-line PR: [TEST RESULT PENDING]
- Did we hit the allowance mid-task? [TEST RESULT PENDING]
What it actually cost
Test task: BRIEF v1 — small browser game (title screen, score, restart)
| Tool / plan | Sticker price | Credits used | Dollars spent | Minutes | Bugs fixed by hand |
|---|---|---|---|---|---|
| GitHub Copilot Pro — default model, agent mode | $10/mo ($15 credits) | [TEST RESULT PENDING] | [TEST RESULT PENDING] | [TEST RESULT PENDING] | [TEST RESULT PENDING] |
| GitHub Copilot Pro — lightweight model, agent mode | $10/mo ($15 credits) | [TEST RESULT PENDING] | [TEST RESULT PENDING] | [TEST RESULT PENDING] | [TEST RESULT PENDING] |
Who Pro is probably for (pricing only, not yet tested)
- Leaning yes: people who mostly want autocomplete in their editor. Completions are unlimited on Pro and don't touch the credits.
- Unknown until tested: heavy agent-mode users. Whether 1,500 credits lasts a month depends on per-task cost, which GitHub doesn't publish. [TEST RESULT PENDING]
Comparing options at a similar price? Read Cursor vs Claude Code vs GitHub Copilot, or our review of Cursor's $8 Plus plan.
Verdict pending
[TEST RESULT PENDING] — no verdict until the hands-on test is done. We don't guess.
Sources
Prices and plan details checked Nov 1, 2026. Prices change often; if you spot a change, tell us and we'll update.
- GitHub Copilot — Plans & pricing — plan prices, AI credit allowances, FAQ on credits and budgets
- TechCrunch — “What a joke”: GitHub Copilot's new token-based billing spurs consternation among devs (May 30, 2026)
- The Register — GitHub Copilot users threaten exit as metered billing kicks in (Jun 2, 2026)
- PlanDiff — secondary: June 1, 2026 switch to token-based billing