\
  The most prestigious law school admissions discussion board in the world.
BackRefresh Options Favorite

Funny how open weight models can be evaluated and tested but OpenAI can't

since they swap out models on the back end or bump you down ...
https://i.imgur.com/YbMiOBE.png
  09/13/26
Link to OpenAI cannot be tested
TurboGrafx-67
  09/13/26
(illiterate)
https://i.imgur.com/YbMiOBE.png
  09/13/26
So no link will be forthcoming
TurboGrafx-67
  09/13/26
Did you hear guise? Anthropic promised not to do this no mor...
https://i.imgur.com/YbMiOBE.png
  09/13/26
I see a link but it doesn't match what OP said
TurboGrafx-67
  09/13/26
hi RSF
https://i.imgur.com/YbMiOBE.png
  09/13/26
Link?
TurboGrafx-67
  09/13/26
that's a non-denial
https://i.imgur.com/YbMiOBE.png
  09/13/26
(no link)
TurboGrafx-67
  09/13/26
https://i.imgur.com/hWmaZZO.jpeg
https://i.imgur.com/YbMiOBE.png
  09/13/26
"Some of the safeguards Anthropic decided on were unsur...
https://i.imgur.com/YbMiOBE.png
  09/13/26
GPT-6 Astra Users Say OpenAI's Newest Model Got Dumber. It H...
https://i.imgur.com/YbMiOBE.png
  09/13/26
this thread sure triggered some folks
https://i.imgur.com/YbMiOBE.png
  09/13/26


Poast new message in this thread



Reply Favorite

Date: September 13th, 2026 2:16 PM
Author: https://i.imgur.com/YbMiOBE.png


since they swap out models on the back end or bump you down to lower quants without informing you

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133609)



Reply Favorite

Date: September 13th, 2026 2:20 PM
Author: TurboGrafx-67

Link to OpenAI cannot be tested

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133616)



Reply Favorite

Date: September 13th, 2026 2:28 PM
Author: https://i.imgur.com/YbMiOBE.png


(illiterate)

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133633)



Reply Favorite

Date: September 13th, 2026 2:30 PM
Author: TurboGrafx-67

So no link will be forthcoming

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133636)



Reply Favorite

Date: September 13th, 2026 2:32 PM
Author: https://i.imgur.com/YbMiOBE.png


Did you hear guise? Anthropic promised not to do this no more. Now they TELL you when you're bumped over to an older model

https://i.imgur.com/r0WpQZw.jpeg

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133640)



Reply Favorite

Date: September 13th, 2026 2:33 PM
Author: TurboGrafx-67

I see a link but it doesn't match what OP said

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133642)



Reply Favorite

Date: September 13th, 2026 2:33 PM
Author: https://i.imgur.com/YbMiOBE.png


hi RSF

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133646)



Reply Favorite

Date: September 13th, 2026 2:33 PM
Author: TurboGrafx-67

Link?

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133647)



Reply Favorite

Date: September 13th, 2026 2:34 PM
Author: https://i.imgur.com/YbMiOBE.png


that's a non-denial

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133651)



Reply Favorite

Date: September 13th, 2026 2:36 PM
Author: TurboGrafx-67

(no link)

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133658)



Reply Favorite

Date: September 13th, 2026 2:35 PM
Author: https://i.imgur.com/YbMiOBE.png


https://i.imgur.com/hWmaZZO.jpeg

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133655)



Reply Favorite

Date: September 13th, 2026 2:37 PM
Author: https://i.imgur.com/YbMiOBE.png


"Some of the safeguards Anthropic decided on were unsurprising: The company said it would reroute users who asked questions about cybersecurity, biology, or chemistry to a less capable AI model"

https://www.wired.com/story/anthropic-responds-to-backlash-on-claudes-secret-sabotage-on-ai-research/

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133662)



Reply Favorite

Date: September 13th, 2026 3:23 PM
Author: https://i.imgur.com/YbMiOBE.png


GPT-6 Astra Users Say OpenAI's Newest Model Got Dumber. It Happened Before, Too

A week after launch, complaints are rolling in from users that GPT-6 Astra has been nerfed. OpenAI's last model went through the same cycle in July.

Jose Antonio Lanz

By Jose Antonio Lanz

Edited by Guillermo Jimenez

Sep 12, 2026

https://decrypt.co/378101/gpt-6-astra-openai-model-dumber-nerfed

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133765)



Reply Favorite

Date: September 13th, 2026 4:43 PM
Author: https://i.imgur.com/YbMiOBE.png


this thread sure triggered some folks

(http://www.autoadmit.com/thread.php?thread_id=5903512&forum_id=2.#50133914)