fix: Fix TTFT in AI SDK v6#2120
Merged
Merged
Conversation
| ); | ||
| case "raw": | ||
| return rawValueHasAISDKContent(part.rawValue); | ||
| default: |
There was a problem hiding this comment.
there's a couple more of these we have to handle, like file.
Member
Author
There was a problem hiding this comment.
I think you're right about file, but not about the others. We only want to track chunks that actually stem from models/generation not any sort of bookkeeping/protocol.
# Conflicts: # js/src/instrumentation/plugins/ai-sdk-plugin.ts
Abhijeet Prasad (AbhiPrasad)
approved these changes
Jun 16, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The problem was that we treated literally the first chunk from the stream as the TTFB trigger, however the first chunk is not necessarily (and actually) LLM tokens, but instead bookkeeping stuff so it doesn't make sense to use it for TTFB.
Instead we now wait for the first chunk that can be attributed to LLM output (text-delta, tool-calls, ...)
Fixes #2110