v2.1.175

Source: Anthropic Claude Code Date: 2026-06-12 URL: https://github.com/anthropics/claude-code/releases/tag/v2.1.175

Summary

v2.1.175 ships a single managed-settings addition: enforceAvailableModels. When enabled, the availableModels allowlist now also constrains what the Default model resolves to — if the Default would fall outside the list, it falls back to the first allowed model — and user or project settings can no longer widen a managed list. This closes a gap where model governance applied to explicit picks but not to the implicit Default resolution path.

Implications

• Fleet-correctness / model governance. This is the control-plane release for organizations that have availableModels deployed: previously an admin could lock the picker but leave the Default path as an escape hatch. The fix makes the allowlist a true ceiling. Relevant for any deployment where compliance or cost control depends on model selection being deterministic rather than advisory.
• Coding-agent host war. Managed settings like enforceAvailableModels are only meaningful if the host (Anthropic’s own Teams/Enterprise tier, or a third-party operator using the managed-settings API) can push them. This widens the gap between Claude Code running in managed vs. unmanaged contexts, which is increasingly where the competitive surface of the host war lives — not just model quality but governance depth.