Fetching from the wire…
Public story · 2026-09-20 · high
The billing reaches back to agents built before August 3, including ones sitting in trial and dev environments.
Why now: Microsoft's grace period, which started September 1, ended September 15, and Modern Workspace Pro's post surfaced the change on September 17.
Microsoft turned on usage-based billing for GitHub Copilot harness agents and workflows in Copilot Studio on September 15, per Modern Workspace Pro's post. The change, tracked as message center item MC1461678, ended a grace period that had run since September 1.
Anyone who built or ran a harness agent in that window, or earlier, now owes credits for it.
Both sides of the meter count. Authoring a maker agent spends Copilot Credits, and so does every runtime execution. The billing reaches back to agents built before August 3, including ones sitting in Dev and Trial environments. A team may have spun those up to test the feature and forgotten they existed.
Standard harness agents and Copilot Chat harness agents are exempt. The charge is specific to the GitHub Copilot harness inside Copilot Studio.
A trial environment used to mean no bill. This notice erases that assumption for one harness type. A team that never got budget approval for Copilot Credits can end up with a live charge attached to a prototype nobody remembers greenlighting.
Each link below shares sources, entities, or timing with this story.
Anthropic's new permanent level indexes to 125 against a 100 baseline, but subscribers have been running at 150 since a temporary 50% boost announced May 13 was extended past July 13, July 19, August 31 and September 13. Both "25% up" and "17% down" are true from different ref...
A spec is a press release until someone who didn't write it implements it. GitHub made Agent Plugins 1.0 generally available on August 12 across VS Code, Copilot CLI, the Copilot SDK, and the Copilot app on all plans. The spec, published August 6, was co-authored by AWS, Anysp...
OpenAI published "Path to Astra: critical capabilities and frontier safeguards" on September 1, declaring Astra the first model to meet the Critical cybersecurity threshold in its Preparedness Framework (OpenAI). Critical, in their own definition, means the model can find and...
An open-weight Chinese frontier model is now a dropdown option in Microsoft's coding product. That happened before anyone finished characterizing what the model does. GitHub's changelog dated August 6 makes Kimi K3 generally available across Copilot Pro, Pro+, Max, Business an...
Six clients. One manifest. Zero vendor lock. Vercel published Agent Plugins 1.0.0 on August 6, an openly licensed spec that bundles Agent Skills and MCP servers behind a single portable manifest. The shape is deliberately boring: a plugin.json requiring only schemaVersion and...
Microsoft added model pinning and hiding plus a comparison view running the same prompt against two models, along with plan consumption visibility and premium model management so teams see quota burn before hitting it. Custom agents can be shared across an organization instead...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.