1.2026.293 is genuinely one of the best ChatGPT updates I’ve seen in months!
First off, READ ALOUD IS BACK! I genuinely couldn’t be happier about this. I’ve been begging for this feature to return for months!
I have dyslexia and rely on Read Aloud for practically every response, especially when using ChatGPT to read stories. Moving it into the overflow menu was an incredibly frustrating accessibility regression that added unnecessary friction to something I use constantly.
Having it back in the main toolbar, with the playback controls working properly, is absolutely fantastic!
Secondly, Reactions appears to be gone! I can’t find it anywhere, and I’m genuinely grateful. It was an addition I never wanted, and I’m very happy to see the interface return to something cleaner and more familiar.
These two changes have genuinely made the app feel like the ChatGPT I fell in love with months ago. Seeing them restored absolutely made my day!
To anyone at OpenAI who might read this: PLEASE don’t change these two things again!
Read Aloud is exactly where it belongs, and I really appreciate the interface without Reactions. This is how the app worked when I first started using it, and I loved it that way.
However, there are still two major problems I’d really like to see addressed.
1. The model picker is still broken.
GPT-6 keeps becoming my default regardless of what I select. I want to use GPT-5.6 Sol, but my selection doesn’t persist, forcing me to repeatedly change models.
Please fix model selection so users can reliably keep their preferred model selected.
2. GPT-6 is a substantial regression for my use case.
I’ll absolutely give GPT-6 credit where it’s due. It’s faster, and I’ve found it better for coding.
Unfortunately, my experience with creative writing has been significantly worse than with GPT-5.6.
I use ChatGPT to generate long-form stories that I read and listen to, with full creative control over the characters, tone, and direction. These stories often involve dark, gritty, mature themes.
GPT-5.6 Sol handles this remarkably well. It follows my storytelling instructions, maintains the tone I’m looking for, and generally respects the creative direction I provide.
GPT-6, by comparison, frequently takes instructions too literally, becomes unnecessarily restrictive with non-graphic mature content, and disregards important storytelling requirements. It often feels like I’m fighting the model instead of collaborating with it.
This has made GPT-6 extremely frustrating for my particular workflow.
Please don’t retire GPT-5.6 until GPT-6 can offer a genuinely comparable creative-writing experience.
I want to keep using ChatGPT. I’ve invested a tremendous amount of time into my stories and workflows here, and I don’t want to lose the model that makes them enjoyable.
My biggest requests are simple:
- Keep Read Aloud in the main toolbar with its full playback controls.
- Keep the interface free of Reactions, or at least give users control over them.
- Fix the model picker so GPT-5.6 can remain selected.
- Improve GPT-6’s instruction-following and creative-writing capabilities.
- Preserve access to GPT-5.6 until those issues are resolved.
I genuinely appreciate the improvements in 1.2026.293. It’s the first update in a long time that has made me this happy to use the app again.
Please keep moving in this direction!