Select users may already have access to the model.
Nice, sounds like next thursday is gonna be big: GPT-5.6 release incoming
If you're wondering how people on your timeline seem to have access to GPT-5.6 Pro, it's now being stealth tested when 5.5 Pro is selected in ChatGPT (*at least for some Pro accounts)!
I have also been told next Thursday is the planned launch date for 5.6
Many users anticipate the GPT-5.6 Pro launch with excitement while others criticize OpenAI for delays, performance regressions like slowness and bugs, and failing to release faster against competitors.
No Digg Deeper questions have been answered for this story yet.
Early stealth test flights are more common in this industry than most people realize. OpenAI has always done it. You may have access to GPT-5.6 right now.
If you're wondering how people on your timeline seem to have access to GPT-5.6 Pro, it's now being stealth tested when 5.5 Pro is selected in ChatGPT (*at least for some Pro accounts)!
I have also been told next Thursday is the planned launch date for 5.6
@AndrewCurran_ i don’t think any queries were routed to gpt-5.6. in fact, gpt-5.5 has been secretly routing me to medium effort without me selecting it
@haider1 Yes it has sometimes rerouted me down to Instant in the last few weeks with Extended thinking selected, very annoying. I understand they are compute constrained but please just be honest about this stuff.
@AndrewCurran_ What would be a good test to distinguish
@vladg_tw A hard prompt or question that you got a particularly good response to so you can compare.
@the_crazydev That would be crazy. Looking forward to next week
@kimmonismus They will make sure to release it right after Fable is back online, just to take momentum from Anthropic...😅
@kimmonismus our AI community just waiting gpt 5.6 release
@AndrewCurran_ True
@AndrewCurran_ they should stop chickening and release the model while anthropic is banned else they are becoming irrelevant increasingly
how dare you
it was a dream that revealed to the alchemists behind the scenes that they must alter the post-training / RL mix by 0.01x thereby increasing head-to-head win rates by 1.4%
yet here you are, doubting their divine taste and calibration
GPT 5.6 will be released next Thursday on June 25
They have chosen the weaker GPT 5.6 checkpoint Kindle-Alpha instead of Kepler-Alpha
@Pauliespasta That sounds fantastic. What is your expertise field?
@kimmonismus Will still not even beat Opus 4.8
@kimmonismus They dropped the ball by not releasing it now when Fable is chained up imo. Was looking forward to using it till Fable got back.
@kimmonismus Just had it auto-routed to 5.6 Pro. It thought for about 3x the average 5.5 Pro thinking time, and the result was noticeably better. Not sure it’s Fable-class, but it was getting close imo. But it was only one prompt in my expertise field.
@kimmonismus To be honest, it's terrible. 5.6 has become as slow as 5.4, the 3D has been slightly improved, the UI has not been fixed.
UI, I can forgive. but returning to the slow speed of GPT 5.4 is very bad.
@kimmonismus @kimmonismus hope it fixes that annoying bug from last version. been driving me nuts lol
@AndrewCurran_ Oh so that's why the models sometimes randomly lock in.
@kimmonismus stealth testing days are always the funniest timeline chaos
@kimmonismus really curious to see how they handle the marketing / benchmarks, after the fable fiasco
will they ignore anthropic models like they did in the LifeScience charts?
