Top
Best
New

Posted by mudkipdev 1 day ago

GPT-5.4(openai.com)
https://openai.com/index/gpt-5-4-thinking-system-card/

https://x.com/OpenAI/status/2029620619743219811

913 points | 721 commentspage 10
freedomben 22 hours ago|
> When toggled on, /fast mode in Codex delivers up to 1.5x faster token velocity with GPT‑5.4. It’s the same model and the same intelligence, just faster.

I hate these blog posts sometimes. Surely there's got to be some tradeoff. Or have we finally arrived at the world's first "free lunch"? Otherwise why not make /fast always active with no mention and no way to turn it off?

cheevly 18 hours ago|
Try improving your attention to detail / reading skills.
ltbarcly3 23 hours ago||
Not a single comparison between 5.4 and Gemini or Claude. OpenAI continues to fall further behind.
world2vec 1 day ago||
Benchmarks barely improved it seems
melbourne_mat 1 day ago||
Quick: let's release something new that gives the appearance that we're still relevant
tmpz22 1 day ago||
Does this improve Tomahawk Missile accuracy?
ch4s3 1 day ago|
They're already accurate within 5-10m at Mach 0.74 after traveling 2k+ km. Its 5m long so it seems pretty accurate. How much more could you expect?
mikkupikku 1 day ago|||
You could definitely do better than that with image recognition for terminal guidance. But I would assume those published accuracy numbers are very conservative anyway..
keithnz 1 day ago|||
I think for LLM like Open AI, it wouldn't be about hitting the target but target selection. Target selection is probably the most likely thing that won't be accurate
ulfw 19 hours ago||
So desperate how they're bumping out these 'updates'
brcmthrowaway 1 day ago||
How much of LLM improvement comes from regular ChatGPT usage these days?
gigatexal 1 day ago||
Is it any good at coding?
throwaway5752 1 day ago|
Does this model autonomously kill people without human approval or perform domestic surveillance of US citizens?
More comments...