The most prestigious law school admissions discussion board in the world.
|
 |
 
|
Even Grok 4 is still WAY BEHIND gpt-5 pro
| metaepistemology is trans | 11/14/25 |
| Mainlining the $ecret Truth of the Univer$e | 11/14/25 |
| metaepistemology is trans | 11/14/25 |
| Mainlining the $ecret Truth of the Univer$e | 11/14/25 |
Poast new message in this thread
Date: November 14th, 2025 3:17 PM Author: metaepistemology is trans Haven't used 5.1 enough to decide on that one but i know for sure from using grok 4 and gpt 5 pro extensively that although grok 4 wins in certain niche tasks and also especially in mathematical precision in many contexts, it still lags pretty substantially behind gpt-5 pro and then the gap after grok 4 is GIGANTIC (for models from anthropic, google, perplexity, deep seek etc) (http://www.autoadmit.com/thread.php?thread_id=5798050&forum_id=2Reputation#49430832) |
Date: November 14th, 2025 3:18 PM Author: Mainlining the $ecret Truth of the Univer$e (One Year Performance 1978-1979 (Cage Piece) (Awfully coy u are)) Gemini 2.5 Pro & Claude Max 5x & NotebookLM (unified via Supermemory.ai (MCP)) are the best for non-developer tasks My Mahchine™ my friend
|
 |
Date: November 14th, 2025 3:23 PM Author: metaepistemology is trans I believe that. I like gemini a lot for its clear writing and accurate summarizing of archival knowledge. It definitely understands everything you give it too. It is just limited in giving critique and usable output for developer tasks and math. Also Claude is good for giving adversarial critiques on more philosophical material, but it depends on context because to me it seems they spent too much time on safety stuff that biases it sometimes. I'm somewhat trolling when I talk gpt-5 pro up so much because my brain is broken from poasting on xo for so long so I can't help but be a provacateur (http://www.autoadmit.com/thread.php?thread_id=5798050&forum_id=2Reputation#49430859) |
|
Date: November 14th, 2025 3:24 PM Author: Mainlining the $ecret Truth of the Univer$e (One Year Performance 1978-1979 (Cage Piece) (Awfully coy u are)) Gemini Pro 2.5 Is significantly less censored on most things compared to ChatGPT Plus & Claude (although you can push the latter two - a bit - with clever prompting) (http://www.autoadmit.com/thread.php?thread_id=5798050&forum_id=2Reputation#49430861) |