In Standard Voice Mode, GPT-6 is all I get, regardless of the model I select in settings. Even worse, since the recent update, voice conversations now switch to Instant with no option to change it.
I’m cancelling my subscription because hands-free access to full model intelligence is essential to my workflow.
GPT-6 is already a significant downgrade from GPT-5.6 Sol in reasoning quality, analytical depth, and contextual fidelity. The Instant version is even worse: a fluent text generator without substance.
It seems that OpenAI is aggressively optimizing for cost efficiency, so they’re pushing “faster”model on users, so GPT-6 keeps appearing even when I explicitly select GPT-5.6 Sol - effectively overriding my model preference.
Contrary to OpenAI's claims that GPT-6 is faster because it’s more capable, my experience is very different: the model consistently avoids doing the analytical work by replacing an actual task with simplistic approximations, ignoring the relevant input data at its convenience to default to general pattern matching even during tasks where context-sensitive reasoning is required.
It responds quickly with long, colorful PDFs, but often fails to address the task that was actually given. Ultimately, this speed is achieved by sacrificing analytical substance and contextual integrity - a clear degradation of the product I’m paying for.
