Anthropic's patched Claude models stay live on three platforms while vulnerable versions remain in production

Anthropic patched the jailbreak affecting Claude Opus 4.6, Opus 3, and Haiku 4.5 in Opus 4.7 onward. Yet all three vulnerable models remain available through the Anthropic API, Azure Foundry, and Amazon Bedrock as of August 2026. The gap between patch and remediation leaves known-vulnerable weights deployed across multiple inference platforms that Anthropic cannot unilaterally update.