Google AI Mode: Availability & Limitations

Availability

Google’s chatbot “AI Mode” is available in the EU - at least partially: Germany works, but one commenter on X reported hitting access restriction from France. The official list of supported countries is published here. It can be reached either through google.com/ai on the “AI Mode” (🇩🇪 “KI-Modus”) button on google.com straight. It’s free of cost but requires a Google account.

Limitations

Philipp Schmid shared this feature list on X:

What it can do:

• - Queries supported via text, voice, images, or PDF uploads.

• - “query fan-out” searches multiple subtopics simultaneously.

• - Resume previous search via AI Mode history.

• - Can use Gemini 2.5 Pro model for advanced reasoning.

• - Allows conversational follow-up questions to refine searches.

However, AI Mode is officially positioned as a Search product:

AI Mode is Google’s most powerful AI search experience […] It divides your question into subtopics and searches for each one simultaneously. This way, AI Mode can explore the web to find even more relevant content that matches your question.

The PDF Upload feature advertised did not immediately work for me: PDF files were grayed out in the file selector. The reason for this is that the “attach file” button is absent in the German version of the user interface. Switching to English helps:

As a workaround to keep the German interface, the trick is to hack the HTML and either amend the list of permitted file types or remove the list altogether, thus allowing any upload: target the file upload button, right-click, choose “Inspect Element” from the popup-menu, find the <input> nearby, and modify the list - e.g. by appending “, application/pdf”:

PDF Analysis currently seems limited to roughly 45 pages. With a longer document, I received this error message and facts were only included up to that limit:

Current limitations only allow part of the document to be used for this answer.

Deleting the pages that had already been analyzed to re-uploading the document with a follow-up prompt was possible, though.

Gemini 2.5 Pro support may be restricted to paying subscribers. From the Support page:

Google AI Pro and AI Ultra subscribers can experience the 2.5 Pro model in Deep Search or select the 2.5 Pro model from the drop-down menu in the AI Mode tab. The default model in AI Mode will continue to be helpful for fast, all-around assistance on most questions.

The advertised Voice support is restricted to live input, however (and cannot be enabled with the file attachment workaround above): when I uploaded an MP3 file, the spinner never finished and the submit button was never enabled.

AI Mode History seems flaky: it doesn’t always save properly even in the same browser session, syncs only partially across devices, and sessions that include a PDF file attached by applying the workaround never seem to get saved.

Conclusion

A powerful model being available for free is certainly nice - the Artificial Analysis Index puts Gemini 2.5 Flash in the same neighbourhood as Anthropic’s Haiku 4.5, GLM 4.6 and DeepSeek V3.2 Exp - and above GPT-5-Chat (not to be confused with GPT-5 Thinking). However, current user interface woes hold AI Mode back - and its positioning as a search product raises questions about its prolonged usefulness for e.g. PDF analysis.