GitHub Copilot code review's Lite and Balanced effort levels โ renamed from the preview's Low and Medium โ are now generally available, letting teams match review depth to pull request complexity.
GitHub's Copilot impact dashboard now includes an ROI section that compares Copilot spend per developer against pull request output, with an adjustable salary selector for modeling scenarios.
GitHub's Copilot usage metrics API adds an optional totals_by_3rd_party_agent array that reports interaction and session counts for agent apps like Claude and Codex running inside GitHub.
GitHub's August 3 weekly Copilot roundup covers new CLI session tools like /worktree and /rewind, VS Code 1.132's multilingual dictation and hybrid Markdown diff editor, and Auto mode model visibility in the Copilot app.
GitHub has resumed rolling out Kimi K3 in GitHub Copilot after briefly pausing it to mitigate a GitHub Actions incident. The model bills at $3/1M input tokens, $15/1M output tokens, and $0.30/1M cached input tokens, and is off by default for Business and Enterprise.
Enterprise owners can now centrally allow or deny specific Model Context Protocol (MCP) servers across GitHub Copilot clients, using new allowedMcpServers and deniedMcpServers keys in managed-settings.json.
Starting August 4, 2026, GitHub Spark no longer accepts new users or new app creation. Existing users can export their apps until August 31, 2026, after which the service winds down.
GitHub has retired the Copilot Billing Preview app as of August 4, 2026, since the standard billing settings now offer more detailed views, including user-level budgets and cost centers.
GitHub Copilot cloud agent now lets users set a reasoning level for models that support it when delegating a task, trading off response quality against token and credit consumption.