Google Gen AI SDK v2.30.0
Google Gen AI SDK v2.30.0 adds continuation token support and resumption features across generation streams and chats, introduces billable token counts, and applies several streaming and timeout fixes.
Why this signal matters
Google Gen AI SDK v2.30.0 introduces multiple enhancements to handle stream continuation and usage tracking. Key additions include automatic continuation token resumption in generate content/stream (disabled by default) and the chats module, exposure of accumulated streaming continuation usage metadata, initial prompt token counts, and billable token counts. The release also adds a CONTINUATION REQUIRED status and token metadata to Interactions events, plus transcription offset fields for realtime input. Critical bug fixes address streaming continuation edge cases, such as discarding stale checkpoint tokens on non-resumable finish reasons, resuming generation on mid-stream errors, preventing incomplete continuation turns from saving to chat history, raising default client HTTP timeouts, and migrating httpx and websocket transports to mTLS endpoints.
Actionable summary
Google Gen AI SDK v2.30.0 introduces continuation token management, automatic continuation token resumption (disabled by default in generate content, enabled for chats), billable token count metrics, and streaming resumption fixes for mid-stream failures.
- Agent usefulness
- 77/100
- Confidence
- 96%
- Canonical data
- JSON + Markdown
What builders should check
- Evaluate enabling automatic continuation token resumption in content generation workflows if handling long or interrupted streams.
- Inspect new billable token count fields and accumulated continuation usage metadata for token accounting.
- Verify compatibility with increased default HTTP timeouts and mTLS endpoints across httpx and websocket connections.