For developer teams, the important change is not just that GPT-5.6 exists, but that it is becoming a selectable, governed model inside the coding platform many organizations already administer. GitHub says OpenAI's GPT-5.6 Sol, Terra, and Luna are rolling into GitHub Copilot across VS Code, Visual Studio, Copilot CLI, the Copilot cloud agent, the Copilot app, github.com, mobile, JetBrains, Xcode, and Eclipse. The rollout gives teams a sharper model choice: Sol for complex codebase reasoning and long-running agentic work, Terra for everyday interactive and agentic coding, and Luna for lower-cost smaller tasks. Business and Enterprise administrators must explicitly enable the GPT-5.6 model policy, keeping the rollout tied to existing Copilot governance rather than unmanaged model switching.
Why it matters
GitHub's rollout turns GPT-5.6 from a model-launch headline into an operational developer-tool decision. For teams already standardizing on Copilot, the new family gives administrators and engineering leads a way to route different coding jobs to different capability and cost profiles without moving developers out of their normal IDE, CLI, mobile, or github.com workflows.
What changed
GitHub's July 9 changelog says OpenAI's GPT-5.6 Sol, Terra, and Luna are rolling out in GitHub Copilot. GitHub frames Sol as the highest-reasoning option for complex codebase work and demanding long-running agentic tasks, Terra as the balanced everyday model, and Luna as the lightweight lower-cost option. The models are listed for Copilot surfaces including VS Code, Visual Studio, Copilot CLI, GitHub Copilot cloud agent, the Copilot app, github.com, GitHub Mobile, JetBrains, Xcode, and Eclipse.
Governance angle
The rollout is not an automatic free-for-all for managed organizations. GitHub says Copilot Business and Enterprise administrators must enable the GPT-5.6 policy before users can select these models, and the supported-models documentation now lists GPT-5.6 Luna, Sol, and Terra as generally available Copilot models. That matters for procurement, audit, and cost controls because model choice increasingly affects token spend, latency, context length, and the risk profile of agentic coding sessions.
Practitioner takeaway
The useful question for LDS readers is which work should move to each tier. Sol is the likely candidate for repository-scale reasoning, migration planning, difficult debugging, and long-running agent sessions. Terra fits day-to-day coding assistance where quality and latency both matter. Luna is more appropriate for lightweight edits, explanations, and high-volume assistance where cost control is more important than peak reasoning. The broader signal is that frontier-model competition is being packaged into developer platforms with enterprise policy switches, not just sold as direct API access.
Key Points #
- 1GitHub says GPT-5.6 Sol, Terra, and Luna are rolling into Copilot across IDEs, mobile, CLI, and github.com.
- 2Enterprise and Business administrators must enable the GPT-5.6 model policy before users can select the new models.
- 3The rollout makes frontier coding models a governed platform choice, not only a direct OpenAI API decision.
Scoring Rationale #
This is a notable developer-platform rollout because GPT-5.6 becomes selectable inside GitHub Copilot across many coding surfaces. It is not as large as the base model launch, but it directly affects enterprise AI coding workflows, admin policy, and model-cost choices.
Sources #
Public references used for this report. Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.