Repository navigation
@ai-sdk/google-vertex@5.0.111
·
7 commits
to main
since this release
Patch Changes
- bdbeba3: feat(google): add
gemini-4-argonmodel id and long decode support forgenerateContent: requests that reach the per-request cap (CONTINUATION) are continued automatically with usage summed across requests and the first request's prompt token count exposed asproviderMetadata.google.initialPromptTokenCount, and thecontinuationTokenprovider option / provider metadata resume responses that used up theirmaxOutputTokensbudget.generateTextwith the Interactions API now pollsbackground: truemodel calls until they finish. - Updated dependencies [bdbeba3]
- @ai-sdk/google@4.0.95