- Who it helps
- Developers using streamed array output in AI SDK applications
- Availability
- November 11 release notes list ai 6.0.330 and 5.0.289. The newer branch covers Output.array and streamObject; the older branch covers streamObject array output.
When the fix matters
Applications streaming many structured elements may repeatedly revisit earlier data. The patch changes that processing strategy; its actual benefit depends on array size and chunking.
Verify in your application
Compare the installed branch with the relevant release and replay a representative long-array response. Check output validity, completion behavior and CPU time before assuming the patch resolves your own slowdown.
What you can do
- Read the version-specific patch and compare it with your installed branch. Open reference ↗
Where the evidence stops
Release notes checked November 11. No package installed and no performance or compatibility test executed.
Sources read. Public source text checked November 11, 2026 UTC. No third-party project or complete workflow executed.
Sources & further reading
GitHub · implementation
AI SDK 6.0.330 release ↗
Version-specific release read: incremental parsing fix for Output.array and streamObject. No execution performed.
Vercel · page checked 2026-11-11 · Published 2026-11-11
GitHub · implementation
AI SDK November 11 patch releases ↗
Versions 6.0.330 and 5.0.289 fix repeated long-array parsing. Patch not installed or benchmarked.
Vercel · page checked 2026-11-11 · Published 2026-11-11
Coverage note: Release notes checked November 11. No package installed and no performance or compatibility test executed.
The record over time (1)
Dated source checks and published explanations. An observation date can differ from the original release date.
- · Added to Current
Vercel’s AI SDK patches a CPU-intensive path when streaming long arrays. Completed elements are parsed and validated incrementally instead of reprocessing the accumulated array on each text chunk.