News for hackers. Stuff that matters.
Re:GPT-5 by joshmlewis on Thursday August 07, 2025 @05:28PM (#44830601)
It's a really good model from my testing so far. You can see the difference in how it tries to use tools to the greatest extent when answering a question, especially compared to 4.1 and o3. In this example it used 6! tool calls in the first response to try and collect as much info as possible.

https://promptslice.com/share/b-2ap_rfjeJgIQsG

Re:GPT-5 by hollownobody on Thursday August 07, 2025 @05:48PM (#44830787)
720 tool calls? Amazing!
Re:GPT-5 by joshmlewis on Thursday August 07, 2025 @05:57PM (#44830876)
Where'd you get 720 from?
Re:GPT-5 by brian626 on Thursday August 07, 2025 @06:03PM (#44830940)
Math pun… 6! = Factorial(6) = 720
Re:GPT-5 by joshmlewis on Thursday August 07, 2025 @06:05PM (#44830952)
Whoosh, it went right over my head.
Re:GPT-5 by terhechte on Thursday August 07, 2025 @06:01PM (#44830920)
the _6!_
Re:GPT-5 by Zone3513 on Thursday August 07, 2025 @06:01PM (#44830916)
That movie doesn't even exist. There is no Thunder Run from 2025.
Re:GPT-5 by joshmlewis on Thursday August 07, 2025 @06:02PM (#44830933)
The data is made up, the point is to see how models respond to the same input / scenario. You're able to create whatever tools you want and import real data or it'll generate fake tool responses for you based on the prompt and tool definition.

Disclaimer: I made PromptSlice for creating and comparing prompts, tools, and models.

Re:GPT-5 by mustaphah on Friday August 08, 2025 @04:49AM (#44834900)
Is there any value in using XML elements to guide the model instead of simple text (e.g., "Recommendation criteria:")?
Re:GPT-5 by flexagoon on Friday August 08, 2025 @05:07AM (#44834985)
XML tags generally help models understand prompts better. That's how most official system prompts are written and what the Anthropic prompting guide says.