databricks Platform · Updated Jul 8, 2026

Google Gemini 2.5 Flash Retirement

Transitioning to newer models for enhanced performance.

What it does

Google Gemini 2.5 Flash will be officially retired on October 2, 2026, meaning it will no longer be accessible for use in any workloads. Users currently utilizing this model are encouraged to migrate to either Gemini 3.1 Pro or Gemini 3.5 Flash, which are recommended replacements. The retirement process follows Databricks' model lifecycle policy, which includes phases of update, deprecation, and retirement to ensure users are informed and can transition smoothly.

Why it matters

This retirement is crucial for maintaining access to state-of-the-art AI capabilities, as older models may not support the latest advancements in machine learning. Transitioning to the newer models will enhance performance and reliability, ensuring that organizations can leverage the most effective tools for their workloads. This change also helps mitigate risks associated with using outdated technology, which can lead to inefficiencies and increased operational costs.

Key capabilities

Model Lifecycle Management

Databricks manages models through a structured lifecycle from update to retirement.

Migration Guidance

Users are provided with clear recommendations for transitioning to replacement models.

Deprecation Notifications

Customers receive advance notifications about model deprecation and retirement timelines.

Access Control

Only workspaces with existing workloads can access deprecated models until retirement.

Recommended use cases

Organizations using Gemini 2.5 Flash for AI applications.Teams needing to upgrade to maintain performance and support.Developers looking to leverage the latest model capabilities.

Limitations · remember

!Once retired, the model will no longer be accessible.

!Existing workloads using the retired model will stop functioning.

!Users must migrate to recommended models before the retirement date.

Architecture

Foundation Model APIs → Model Management → Workload Transition → Model Replacement

Get started

Review Current Usage

Identify any workloads currently using Google Gemini 2.5 Flash.

Select Replacement Model

Choose between Gemini 3.1 Pro or Gemini 3.5 Flash as the new model.

Plan Migration

Develop a migration plan to transition workloads to the new model.

Implement Changes

Update your applications to use the selected replacement model.

Test Workloads

Verify that workloads function correctly with the new model before the retirement date.

Key takeaway

Transitioning from Google Gemini 2.5 Flash to newer models is essential for maintaining cutting-edge AI capabilities.

2026-07-08-google-gemini-2-5-flash-is-scheduled-for-retirement Generated by the Databricks Updates Agent · browse all updates