The AI community is buzzing with the news that a preview of Gemini 3.0 Pro is now accessible through Google Vertex AI, albeit with some significant limitations. The model, identified as gemini-3-pro-preview-11-2025, is only available in a few select regions and to a limited number of accounts.

For those who don't have direct access through Google Vertex AI, the community has discovered a few alternative methods via Chinese API aggregators. These platforms provide access to the model, though with their own set of rules and limitations.

Here are a couple of the aggregators where the model has been spotted:

It's likely that more aggregators will start listing the model as availability expands. If you find other sources, sharing is caring!

It's important to set expectations: the Gemini 3.0 Pro preview is currently highly unstable. Users have reported frequent request failures and artificial limitations on context and output tokens. This suggests that the model is in a very early stage of testing and is not yet ready for production workloads.

If you're a developer or researcher eager to get a first look at the capabilities of Gemini 3.0 Pro, these aggregators offer a tantalizing glimpse. However, for any practical applications, it's best to wait for a more stable and official release from Google.

We'll be keeping a close eye on the situation and will provide updates as more information becomes available. Stay tuned!