The most prestigious law school admissions discussion board in the world.
|
 |
 
|
ChatGPT Plus 5.2 is *almost* as good as Claude's Opus 4.5
Poast new message in this thread
Date: January 6th, 2026 11:18 PM Author: Cordovan house (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49568112) |
Date: January 7th, 2026 6:17 AM Author: Cerebral Bright Kitchen 5.2 is significantly better IMO. In terms of getting sane and accurate answers. Claude is better at writing though. GPT still has a janky style to it’s outputs (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49568388) |
 |
Date: January 7th, 2026 6:23 AM Author: thriller frisky legal warrant Chatgpt writes
Like Jordan peterson
Withdrawing from benzos on
Twitter |
|
Date: January 7th, 2026 8:37 AM Author: Misanthropic salmon codepig Your AI threading is the most retarded shit on xoxo rn (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49568494) |
|
Date: January 7th, 2026 9:40 AM Author: Cordovan house Explain? (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49568617) |
|
Date: January 7th, 2026 10:00 AM Author: aquamarine hell international law enforcement agency it still made up a case when i asked it a fairly specific question. Made up case name, cite, what it held, etc. (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49568677) |
|
Date: January 7th, 2026 10:25 AM Author: shivering sapphire library pistol you need to ground it with a claude skill and integrate exa search etc (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49568719) |
|
Date: January 7th, 2026 10:27 AM Author: Cordovan house (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49568722) |
|
Date: January 10th, 2026 2:15 AM Author: Mainlining the $ecret Truth of the Univer$e (One Year Performance 1978-1979 (Cage Piece) (Awfully coy u are)) https://youtu.be/2WTFuSeYJ6o?si=sucecMqeSbI9qQaQ&t=1600 (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49578185) |
|
Date: January 10th, 2026 2:24 AM Author: STEPHEN MILLER (FAGGOTCHIPPER / Hegemon) I think 5.2 is significantly less useful than any prior model. the safety / compliance limits are ultra sensitive and neuter it even with materially benign / open-source information. (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49578231) |
|
Date: January 10th, 2026 2:28 AM Author: Mainlining the $ecret Truth of the Univer$e (One Year Performance 1978-1979 (Cage Piece) (Awfully coy u are)) It's 180 compared to Gemini 3 Pro (and loved Gemini 2.5 Pro) (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49578237) |
|
Date: January 10th, 2026 9:39 PM Author: chopped unc its more aligned with published knowledge than ever. which is good for abstract and highly technical subject matter. horrible for anything highly social, political, uncomfortable or where human bias seeps in (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49579884) |
|
Date: January 14th, 2026 7:46 PM Author: claudeman . (http://www.autoadmit.com/thread.php?thread_id=5818216&forum_id=2#49590139) |