OpenAI released Lockdown Mode to all users to restrict network-enabled capabilities and reduce data exfiltration risks.

Security now takes precedence over utility by disabling live browsing and agent modes to stop prompt injection attacks.

This setting creates a closed environment for sensitive data handling by removing the external connectors that typically enable autonomous operations.

• Jun 06, 2026Anthropic updated prompt caching to automatically read from the longest previously cached prefix when breakpoints are set.
• Jun 06, 2026Anthropic optimized prompt caching support in the Message Batches API to improve cache hit rates.
• Jun 06, 2026Anthropic added a delete endpoint to the Message Batches API for managing batch processing tasks.

This is the world we build in. It moves this fast. We move with it.