【朗報！】GitHub Copilot の新モデルを、管理者の設定変更なしで自動的にユーザーへ有効化できるようになりますGitHub Copilot Enterprise added a 'Default availability for released models'…

GitHub has added a new administrative policy to GitHub Copilot Enterprise called "Default availability for released models," which determines whether newly released AI models become available to users automatically or wait for explicit approval from an administrator. The change matters because it directly affects how quickly development teams can reach the latest models while also shaping how much control an organization keeps over which models run inside its environment.

The policy addresses a friction point that has become more visible as the number of models offered through Copilot has grown. GitHub Copilot no longer depends on a single underlying model. Over recent releases the service has added support for models from multiple providers, and administrators of Enterprise and Business accounts have generally needed to review and enable each new model through the model management settings before their users could select it. That approach gives organizations tight control, but it also means that every model release requires a manual step, and users cannot access a newly launched model until an administrator acts.

Under the "Default availability for released models" setting, an enterprise can decide this default behavior in advance rather than handling each release one at a time. When the policy is configured to make models available by default, newly released models are turned on for users automatically at launch, without an administrator updating the configuration each time. Organizations that prefer a more cautious posture can keep new models disabled by default, preserving a review-before-enable workflow. In practical terms, the policy shifts model rollout from an opt-in-per-model process to a single configurable default that governs future releases.

That distinction is meaningful for teams with compliance, security, or cost considerations. Some organizations want to evaluate a model's behavior, data handling, and performance before exposing it to developers, and for them the manual gate remains valuable. Others prioritize giving engineers immediate access to the newest capabilities and view the per-model approval step as an unnecessary delay. The new policy appears designed to let each organization pick the trade-off that fits its governance model, rather than imposing one default on everyone.

It helps to place this within the broader set of administrative controls that GitHub already exposes for Copilot. Enterprise and Business plans include policies covering areas such as whether suggestions matching public code are allowed, content exclusion rules that keep specified files or repositories out of Copilot's context, and controls over which Copilot features are enabled for members. Model management sits alongside these as part of the same governance surface, and the addition of a default-availability policy is consistent with GitHub's pattern of turning individual, repetitive administrative decisions into settings that can be established once and applied going forward.

The timing also reflects the pace of change in the underlying AI market. Model providers now ship new versions and entirely new models frequently, and the interval between announcements has shortened considerably. For a platform like Copilot that integrates several of these models, a purely manual enablement process scales poorly when releases arrive in quick succession. A default policy is a reasonable response to that cadence, because it lets administrators set an intent rather than react to each individual launch.

For teams evaluating the change, the practical takeaway is to review how their current model management settings are configured and to confirm which default they want applied to future releases. Organizations that have relied on the manual approval step should verify whether enabling automatic availability aligns with their internal review requirements, since the setting would allow new models to reach users without a further check. Those that want faster access can use the policy to reduce the lag between a model's release and its availability to developers.

The feature was noted through a Copilot-focused community blog rather than framed as a major product launch, and the specifics of configuration and rollout are likely to be documented in GitHub's official administrator materials. As with other Copilot policies, availability and behavior may vary by plan tier, so administrators should confirm the exact options in their own Enterprise management console before relying on the new default.