Gemini-3.7-flash requests are answered as gemini-3.8-flash

Hi all,

We’re seeing that every generateContent request to models/gemini-3.7-flash comes back with “modelVersion”: “gemini-3.8-flash”. We tried different requests with different explicit thinking levels but the results are the same. It however does not happen for other models we tried (3.6 in example).

We reproduced it with:

curl -sS "https://generativelanguage.googleapis.com/v1beta/models/gemini-3.7-flash:generateContent" \
  -H "x-goog-api-key: $KEY" \
  -H 'Content-Type: application/json' \
  -X POST \
  -d '{"contents":[{"parts":[{"text":"Reply with OK."}]}]}' \
  | grep '"modelVersion"'

Expected modelVersion is `gemini-3.7-flash` but response was `gemini-3.8-flash`.

Can we tell if either 3.7 requests are indeed being served by 3.8 or if the model version field is reporting the wrong version?

Hi, we received an email from Google saying all requests to 3.7 are routed automatically to 3.8.

Hello @Joao_Silveira ,

Yes, this is intended behavior. gemini-3.7-flash is deprecated as of October 8, 2026 (and removed from the Google AI Studio model selector), though the API endpoint is not shutting down and automatically routes all requests to gemini-3.8-flash. Please check out the Gemini deprecations page for more details.