GitHub Copilot has two changes worth updating our teaching for: four older models have been retired, and code reviews can now be requested through APIs. No further qualifying change was verified for ChatGPT, Claude, Microsoft Copilot or Gemini in this edition’s window.
New releases
GitHub Copilot. GitHub retired Gemini 3.5 Flash and Gemini 3.6 Flash, recommending Gemini 3.8 Flash; Kimi K2.7 Code, recommending Kimi K3; and Claude Opus 4.7, recommending Claude Opus 5.5. The change applies across Copilot Chat, inline edits, ask and agent modes, and code completions. More
GitHub Copilot. Scripts, workflows and internal tools can now request Copilot code reviews through REST and GraphQL APIs, optionally specifying the effort level for each request. As background, Default has used Balanced since 28 September for new and existing repositories and organisations; explicitly selected Lite settings were retained. More
Elsewhere in AI
Georgia’s election board held an emergency meeting after research showed AI could combine public election records to identify how individual people voted. Read
Australia’s home affairs department ordered federal agencies to review legacy technology after the OpenAI Medicare breach, as the government faces potentially costly cybersecurity improvements. Read
CME Group plans to introduce a futures market for AI compute, allowing businesses to hedge against fluctuations in chip rental prices, the Financial Times reports. Read
US authorities arrested Greg Lui for allegedly smuggling over US$300m of export-controlled computer servers to China, using false paperwork and shipments through third countries. Read