The FinOpsAid blog
Notes on the cost of building on GitHub.
Practical, vendor-neutral writing on GitHub spend — Copilot seats, Actions minutes, Codespaces, storage and the FinOps discipline that keeps them honest. No invented benchmarks, no fluff.
December's Copilot model swaps aren't all price cuts. Two of them double.
GitHub deprecates six Copilot models on December 14, 2026 and names a replacement for five of them. Two replacements are much cheaper, one costs the same, and two cost twice as much per token as what your developers use today. Gemini 4 Carbon's own price also doubles on December 1. Here's the swap-by-swap math from GitHub's price list, and the policy detail that can stop the Claude swap from happening at all.
Read →Copilot's metrics now flag their own blind spots. Read those before you read the rest.
In November 2026 GitHub added four things to Copilot usage metrics. They split out bring-your-own-key and enterprise custom models, measure whether developers act on Copilot code review comments, flag users whose editors send no telemetry, and add chat and the VS Code Agents window to the impact dashboard. Each one sharpens a cost question. One of them shows usage that never reaches your GitHub bill, and another tells you which of your existing charts to distrust.
Read →Which cost center is this developer in? GitHub now answers in one call.
In November 2026 GitHub added a REST endpoint that returns the cost center for a single user, repository, organization or enterprise team, and gave GitHub Apps a permission to manage enterprise Copilot seats. Together they let a bot check who's billed where and who holds a seat, without a personal token tied to an admin. The read side is an easy win. The write side is a seat-removal permission, so scope it like one.
Read →Copilot Automations can now cap their own AI credits. None of them do by default.
On October 28, 2026 GitHub gave Copilot Automations optional AI credit limits per run, day, week and month, alongside 24 new triggers, API triggers and the ability to chain automations together. The limits are opt-in per automation, anyone with write access can create one, and only some of the limits can stop a loop. Agentic Workflows went GA the same day and bill on two meters.
Read →Copilot code review now has a cost dial. Someone turned it up on September 28.
Between September 28 and October 28, 2026, GitHub moved Copilot code review's default effort from Lite to Balanced, added a Max level that uses more AI credits, let organizations choose who pays for a review, and opened reviews to the REST and GraphQL APIs. Each is reasonable on its own. Together they turn review effort into a cost setting that most teams haven't set.
Read →Copilot Fusion picks the model now. You still pay for every one it picks.
GitHub made Copilot Fusion generally available on October 28, 2026. It routes each task across models, and it bills each model it uses at that model's standard rate, so a Critique run can cost several times a single call. Three models also retire on November 16, and this time every replacement is cheaper. One of the cheapest options isn't on GitHub's suggested list at all.
Read →