persistent and critical issue I’m experiencing with the Stream mode functionality within AI Studio, specifically when utilizing the Model 2.5 (“thinking”) architecture
My observations are as follows:
1. Prolonged “Thinking” State with No Output:
I have attempted to generate responses more than 20 times using Model 2.5. In the vast majority of these instances, the AI enters a “thinking” state, indicated by the continuous loading animation (as shown in the attached screenshot). While it appears the model may have completed its internal processing, it remains perpetually in this “thinking” mode without producing any textual or voice output. The system fails to transition from the processing state to a display of the generated response.
2. Incomplete and Abrupt Responses:
On the rare occasions (a very low success rate) when the model does manage to provide a response, the output is consistently truncated or incomplete. The AI frequently stops after only a few words, failing to deliver a full or coherent answer. This behavior severely impacts the usability and reliability of the model.
These issues significantly hinder my ability to effectively use AI Studio and retrieve complete responses from Model 2.5. Please advise on any potential workarounds or a timeline for a fix.