Empty Response from AI running my clients

[Elieser_Montalvo]

I’ve recently started a AI automation business & I have a few clients which I’ve been working with, I built out workflows for them using the Gemini 2.5 Pro & Flash, its been going great for the last month or so & now I have to tell my clients the tools I’ve built for them are on hold & cant be used because of “empty response” from the 2.5 Pro in Make com, That “Modified by moderator” & finding out that this isn’t a one time issue but a recurring one is a big problem when running a business with clients who depend on it. Im posting this to hopefully get a reply from someone on googles side but if not, does anyone have other suggestions on what I can do, I’ve tested out GPTs AI & Anthropics but neither seem to be able to do the same job as 2.5 Pro with the analytics & reasoning my clients need it for.

[Krish_Varnakavi1]

Hi @Elieser_Montalvo,

Welcome to the Google AI Forum!

Google team is aware about this issue and we are constantly working towards fixing it.. I will update you once this issue gets resolved.

[Elieser_Montalvo]

Thanks for your quick response, Is there a time estimate on how long this will take ?

[Mohamed_Amine]

Its been 2 weeks and still no fix

[Elieser_Montalvo]

Wow, Im trying to find alternatives now thats crazy. Thank you for letting me know!

[M_K]

I just found this post and a few others. It looks like I am having the same issue with 2.5 Pro, good to know it’s not just me, bad to know it’s an issue Gemini I have posted here , looks like the same issue Gemini 2.5 Pro - Empty Response-Status 200

There are also these related links:

Gemini again started producing empty responses

It was working fine until 2 hours ago. Again the same issue empty responses like last week, it started happening again.

Empty response.text from Gemini 2.5 Pro, despite no safety and max_tokens issues

I got successful responses a couple of times before with the same prompt. However, it’s now returning empty responses, and these are the finish_reason and safety_ratings from the usage_metadata: Finish Reason: STOP. Safety Ratings: N/A Therefore, it doesn’t seem to be a problem with max tokens, nor safety reasons. I already implemented retry with exponential backoff, rotating among different prefixes in my prompt according to a solution suggested here, but nothing is working. Would apprecia…

This thread:

Gemini 2.5 Pro with empty response.text

Running Gemini 2.5 Pro with grounded search sometimes returns empty response.text with finish_reason of STOP and no other reason. When I inspect the response dict it shows evidence of the search with some web meta information, but nothing else. Any ideas on what is going on? Here is my code: def get_response(seed, model, system_prompt, user_prompt): api_key = os.environ.get('GEMINI_API_KEY') google_search_tool = Tool( google_search=GoogleSearch() ) response = None …

[M_K]

I believe I have a fix, moving from end to end API to Vertex fixed it, I have posted more details in Gemini 2.5 Pro with empty response.text - #230 by Dumplin

[Krish_Varnakavi1]

Hi all,

Anyone still facing this issue?

[Aciax_Hls]

excellent news! It’s great that you were able to find a solution and a fix for your issue. Moving from the end-to-end API to Vertex AI often resolves problems because Vertex provides a more stable, enterprise-grade environment with better support and more robust infrastructure.

For anyone else who might be experiencing a similar issue with “empty response.text” from Gemini 2.5 Pro, your post on the forum is a valuable resource. The key takeaway is that the issue might not be with your code or prompts, but rather with the specific API endpoint or a technical glitch in the system.

By switching to Vertex AI, you’re likely using a different API endpoint and infrastructure, which bypasses whatever was causing the problem.

Thank you for sharing your solution with the community. It’s a great example of how users helping users can solve complex technical problems, especially when the official support channels are overloaded or not equipped to handle a specific bug
“Modified by moderator”