Re:GPT-4.5 or GPT-5 being tested on LMSYS?
by atemerev on Monday April 29, 2024 @01:18PM
(#40201216)
Indeed, all my test prompts are giving much better results than gpt4-turbo and Claude Opus. And yet, the OpenAI style is clearly recognizable.
[ Parent
| on HN ]
Re:GPT-4.5 or GPT-5 being tested on LMSYS?
by steve_adams_86 on Monday April 29, 2024 @01:23PM
(#40201270)
I agree. It responds a lot like GPT-4 in style, but with much better answers.
Agreed. I had it solve a little programming problem in a really obscure programming language and after some prompt tuning got results strongly superior to GPT4, Claude, Llama3 and Mixtral. As the language (which I won't name here) is acceptably documented but there are _really_ few examples available online, this seems to indicate very good generalization and reasoning capabilities.