Greetings. I was looking at the supported regions for the Gemini API in Vertex AI. I have to follow compliance requirements so I was searching for regional endpoints here at this link.
https://cloud.google.com/vertex-ai/generative-ai/docs/learn/locations#google_model_endpoint_locations
I noticed of course the warning about Gemini 1.5 Pro not being available anymore. My project has no prior usage of Gemini 1.5 Pro. This is a problem because on the Canada tab it apparently shows that no future version of Gemini is supported in any Canada region yet, including Montreal? This is quite discouraging and confusing - why is support for this (Gemini 1.5 Pro) being phased out if it is not even ready to be replaced by anything in Montreal?
I would also like to know what other Model as a Service options from Vertex AI model garden I have for my LLM use case because I do not want to deploy an LLM myself and incur node-hour costs.
Thank you very much in advance!
2 Likes
Hi maham1243,
Welcome to Google Cloud Community!
As of the latest update, Gemini 1.5 Pro and Gemini 1.5 Flash models are not available in projects that have no prior usage of these models, including new projects. For details, see Model versions and lifecycle.
Google hasn’t publicly explained why support is being phased out before a replacement is available in Canada, but it likely reflects a combination of infrastructure rollout timelines, demand concentration in other regions, and evolving compliance strategies. I suggest keeping an eye on the Google Cloud release notes and the locations page you linked.
In addition, you may check this documentation for Vertex AI’s Model Garden. Also refer to Vertex AI partner models for Model as a Service, partner models are serverless so there’s no need to provision or manage infrastructure.
Was this helpful? If so, please accept this answer as “Solution”. If you need additional assistance, reply here within 2 business days and I’ll be happy to help.
1 Like
Thank you for your attention to my post! Unfortunately as I have to follow compliance requirements I require support for data residency in Canada and I really cannot use global endpoints. And more unfortunately, no other model offers data residency in any Canadian region.
This is very surprising from the MaaS offerings from Vertex AI which are so diverse that the data residency in Canada is limited to nearly nothing at all. Even the Nvidia V100 and P100 GPUs are not available. I hope Canada moves farther up in the priority list or others like me will be forced to consider alternatives and I don’t want to. For my use case I was prepared to embrace the pricing models of the MaaS offerings - the custom ‘bring your own weights’ pricing model is far out of my budget. Like you said, I will watch the release notes page and hope by Fall they replace Gemini 1.5 FULLY.
1 Like
Apparently gemini-2.5-flash is now available but I am still not able to use it on the northamerica-northeast1 server. Unsure if there is some lag between the documentation and the deployment.
@maham1243 @alexandrecc I believe that Gemini 2.5 Flash has just made available for in-region ML processing in Canada, as shown in documentation](Deployments and endpoints | Generative AI on Vertex AI | Google Cloud)), and the access may only be available via Single-Zone Provisioned Throughput.
1 Like
Wow, thank you so much for the update! Yes, it looks like it’s there but could you please tell me where it’s written that Gemini 2.5 Flash is only available as provisioned throughput because I did not see that on the locations/ page. Where can I find out about that?
Although it’s already looking true because I can only access the model from Iowa and not Montreal.
@maham1243 I believe that the documentation will be updated eventually. Meanwhile, see if you can proceed by connecting to your Google Cloud account representative or following instructions: https://cloud.google.com/vertex-ai/generative-ai/docs/provisioned-throughput/purchase-provisioned-throughput
