Silent model
Google deprecated gemini-3.7-flash on October 8, 2026 and now automatically routes API requests for it to gemini-3.8-flash.
What changes
The October 8, 2026 Gemini API changelog entry says gemini-3.7-flash is deprecated and has been replaced by gemini-3.8-flash, that it is no longer available to select in Google AI Studio, that API requests to gemini-3.7-flash are automatically routed to gemini-3.8-flash, and tells callers to update the model string. Calls keep succeeding, but they are served by a different model. The deprecations page lists gemini-3.7-flash (released August 13, 2026) with shutdown date N/A and gemini-3.8-flash as the replacement, and says the endpoint is not shutting down.
Silent: No error. Behavior, results, or cost change quietly.
What to do
Change the model string from gemini-3.7-flash to gemini-3.8-flash, as Google directs, and review the Gemini 3.8 Flash guide. Your requests are already served by gemini-3.8-flash, so re-check output quality, latency and cost on your evals.