Unable to use gemini-1.5-flash-8b with caching

[Vandivier]

"google-cloud-aiplatform>=1.71.1",
"vertexai>=1.71.1",

note: I can't upgrade google-cloud-aiplatform because PIP tells me it depends on vertexai and there is no higher version of vertexai available

similar to
#4009

except:

1. my issue is particular to 8b
2. my issue is not particular to stable versions (I've tried gemini-1.5-flash-8b-latest in addition to stable gemini-1.5-flash-8b)

[Vandivier]

by my reading, eg here, "The Vertex AI SDK and the Vertex AI Python client library provide similar functionality with different levels of granularity." so i'm surprised if the model support varies

if there is a general pattern that model support rolls out in one package prior to the other, that would be a good note in the docs for consumers