Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

GPT-5.4

openai.com

Post reply on HN
Print view View original

201–210 of 868 posts

Re: GPT-5.4

#201

Post by staticman2 »

bargainbin wrote: post #174
cj wrote: post #105

Earlier quoted context omitted.

I'm sometimes surprised how much detail ChatGPT will go into without giving any dislaimers. I very frequently copy/paste the same prompts into Gemini to compare, and Gemini often flat out refuses to engage while ChatGPT will happily make medical recommendations. I also have a feeling it has to do with my account history and heavy use of project context. It feels like when ChatGPT is overloaded with too much context,…

Anecdotal, but I asked Claude the other day about how to dilute my medication (HCG) and it flat out refused and started lecturing me about abusing drugs. I copy and pasted into ChatGPT, it told me straight away, and then for a laugh said it was actually a magical weight loss drug that I'd bought off the dark web... And it started giving me advice about unregulated weight loss drugs and how to dose them.

If you had created a project with custom instructions and/ or custom style I think you could have gotten Claude to respond the way you wanted just fine.

Re: GPT-5.4

#202

Post by kseniamorph »

Aboutplants wrote: post #31
twtw99 wrote: post #15

If you don't want to click in, easy comparison with other 2 frontier models - https://x.com/OpenAI/status/2029620619743219811?s=20

It seems that all frontier models are basically roughly even at this point. One may be slightly better for certain things but in general I think we are approaching a real level playing field field in terms of ability.

makes sense, but i'd separate two things: models converging in ability vs hitting a fundamental ceiling. what we're probably seeing is the current training recipe plateauing — bigger model, more tokens, same optimizer. that would explain the convergence. but that's not necessarily the architecture being maxed out. would be interesting to see what happens when genuinely new approaches get to frontier scale.

Re: GPT-5.4

#203

Post by satvikpendem »

earth2mars wrote: post #180
ipsum2 wrote: post #111

Earlier quoted context omitted.

The model was released less than an hour ago, and somehow you've been able to form such a strong opinion about it. Impressive!

I am actually super impressed with Codex-5.3 extra high reasoning. Its a drop in replacement (infact better than Claude Opus 4.6. lately claude being super verbose going in circles in getting things resolved). I stopped using claude mostly and having a blast with Codex 5.3. looking forward to 5.4 in codex.

Same, it also helps that it's way cheaper than Opus in VSCode Copilot, where OpenAI models are counted as 1x requests while Opus is 3x, for similar performance (no doubt Microsoft is subsidizing OpenAI models due to their partnership).

Re: GPT-5.4

#204

Post by varispeed »

yanis_t wrote: post #69

These releases are lacking something. Yes, they optimised for benchmarks, but it’s just not all that impressive anymore. It is time for a product, not for a marginally improved model.

The scores increase and as new versions are released they feel more and more dumbed down.

Re: GPT-5.4

#205

Post by satvikpendem »

ipsum2 wrote: post #111
yanis_t wrote: post #69

These releases are lacking something. Yes, they optimised for benchmarks, but it’s just not all that impressive anymore. It is time for a product, not for a marginally improved model.

The model was released less than an hour ago, and somehow you've been able to form such a strong opinion about it. Impressive!

It's more hedonic adaptation, people just aren't as impressed by incremental changes anymore over big leaps. It's the same as another thread yesterday where someone said the new MacBook with the latest processor doesn't excite them anymore, and it's because for most people, most models are good enough and now it's all about applications.

https://news.ycombinator.com/item?id=47232453#47232735

Re: GPT-5.4

#206

Post by landtuna »

surgical_fire wrote: post #134
simianwords wrote: post #107

Earlier quoted context omitted.

No my question was why would I use codex over gpt 5.4

Ahh, good question. I misunderstood you, apologies. There's no mention of pricing, quotas and so on. Perhaps Codex will still be preferable for coding tasks as it is tailored for it? Maybe it is faster to respond? Just speculation on my part. If it becomes redundant to 5.4, I presume it will be sunset. Or maybe they eventually release a Codex 5.4?

5.3 Codex is $1.75/$14, and 5.4 is $2.50/$15.

Re: GPT-5.4

#207

Post by netinstructions »

minimaxir wrote: post #8

The marquee feature is obviously the 1M context window, compared to the ~200k other models support with maybe an extra cost for generations beyond >200k tokens. Per the pricing page, there is no additional cost for tokens beyond 200k: https://openai.com/api/pricing/ Also per pricing, GPT-5.4 ($2.50/M input, $15/M output) is much cheaper than Opus 4.6 ($5/M input, $25/M output) and Opus has a penalty for its beta >200…

People (and also frustratingly LLMs) usually refer to https://openai.com/api/pricing/ which doesn't give the complete picture.

https://developers.openai.com/api/docs/pricing is what I always reference, and it explicitly shows that pricing ($2.50/M input, $15/M output) for tokens under 272k

It is nice that we get 70-72k more tokens before the price goes up (also what does it cost beyond 272k tokens??)

Re: GPT-5.4

#208

Post by NiloCK »

7777777phil wrote: post #151

83% win rate over industry professionals across 44 occupations. I'd believe it on those specific tasks. Near-universal adoption in software still hasn't moved DORA metrics. The model gets better every release. The output doesn't follow. Just had a closer look on those productivity metrics this week: https://philippdubach.com/posts/93-of-developers-use-ai-codi...

This March 2026 blog post is citing a 2025 study based on Sonnet 3.5 and 3.7 usage.

Given that organization who ran the study [1] has a terrifying exponential as their landing page, I think they'd prefer that it's results are interpreted as a snapshot of something moving rather than a constant.

[1] - https://metr.org/

Re: GPT-5.4

#209

Post by PaulHoule »

disqard wrote: post #46
PaulHoule wrote: post #17

Earlier quoted context omitted.

APIs have never been a gift but rather have always been a take-away that lets you do less than you can with the web interface. It’s always been about drinking through a straw, paying NASA prices, and being limited in everything you can do. But people are intimidated by the complexity of writing web crawlers because management has been so traumatized by the cost of making GUI applications that they couldn’t believe ho…

> AI is a threat to the “enshittification economy” because it lets us route around it. This is prescient -- I wonder if the Big Tech entities see it this way. Maybe, even if they do, they're 100% committed to speedrunning the current late-stage-cap wave, and therefore unable to do anything about it.

They are not a single thing.

Google has a good model in the form of Gemini and they might figure they can win the AI race and if the web dies, the web dies. YouTube will still stick around.

Facebook is not going to win the AI race with low I.Q. Llama but Zuck believed their business was cooked around the time it became a real business because their users would eventually age out and get tired of it. If I was him I'd be investing in anything that isn't cybernetic let it be gold bars or MMA studios.

Microsoft? They bought Activision for $69 billion. I just can't explain their behavior rationally but they could do worse than their strategy of "put ChatGPT in front of laggards and hope that some of them rise to the challenge and become slop producers."

Amazon is really a bricks-and-mortar play which has the freedom to invest in bricks-and-mortar because investors don't think they are a bricks-and-mortar play.

Netflix? They're cooked as is all of Hollywood. Hollywood's gatekeeping-industrial strategy of producing as few franchise as possible will crack someday and our media market may wind up looking more like Japan, where somebody can write a low-rent light novel like

https://en.wikipedia.org/wiki/Backstabbed_in_a_Backwater_Dun...

and J.C. Staff makes a terrible anime that convinces 20k Otaku to drop $150 on the light novels and another $150 on the manga (sorry, no way you can make a balanced game based on that premise!) and the cost structure is such that it is profitable.

Re: GPT-5.4

#210

Post by gavinray »

nickysielicki wrote: post #29

can anyone compare the $200/mo codex usage limits with the $200/mo claude usage limits? It’s extremely difficult to get a feel for whether switching between the two is going to result in hitting limits more or less often, and it’s difficult to find discussion online about this. In practice, if I buy $200/mo codex, can I basically run 3 codex instances simultaneously in tmux, like I can with claude code pro max, all d…

I almost never hit my $20 Codex limits, whereas I often hit my Claude limits.
Post reply on HN