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

581–590 of 868 posts

Re: GPT-5.4

#581

Post by bwat49 »

gizmodo59 wrote: post #538
baxtr wrote: post #478

Earlier quoted context omitted.

I picked up Claude today after being away and using only ChatGPT and Gemini for a while. I was pretty impressed with how they’ve improved user experience. If I had to guess, I’d say Anthropic has better product people who put more attention to detail in these areas.

ChatGPT has given more for my 20$ than any other vendor. And that’s not even considering codex which is so good and the limits are much much higher

yeah claude is great... but only if you pay $100-$200 a month

Re: GPT-5.4

#582

Post by nembal »

so it seems each RL step extends into a market! 5.3 was target at coding. 5.4 is target at finance 5.5 is healthcare?

Re: GPT-5.4

#583

Post by creatonez »

> We put a particular focus on improving GPT‑5.4’s ability to create and edit spreadsheets, presentations, and documents.

Nothing infuriates me more than an LLM tool randomly deciding to create docx or xlsx files for no apparent reason. They have to use a random library to create these files, and they constantly screw up API calls and get completely distracted by the sheer size of the scripts they have to write to output a simple documents. These files have terrible accessibility (all paper-like formats do) and end up with way too much formatting. Markdown was chosen as the lingua franca of LLMs for a reason, trying to force it into a totally unsuitable format isn't going to work.

Re: GPT-5.4

#584

Post by abustamam »

thereticent wrote: post #569
abustamam wrote: post #526

Earlier quoted context omitted.

Kinda reminds me of crypto. There are certainly very interesting things happening in the crypto space. But the most visible parts of the crypto universe are the stupid parts (buying PNGs for millions, for example)

Genuinely curious, not being combative...what very interesting things have happened in the crypto space lately?

Oh, I dunno about lately (though I did stumble upon https://a16zcrypto.com/posts/article/big-ideas-things-excite... )

But when I was in the crypto space in 2018, there was a lot of interesting things happening in the smart contract world (like proofs of concepts of issuing NFTs as a digital "deed" to a physical asset like a house).

I don't think any of those novel ideas went anywhere, but it was a fun time to be experimenting.

Re: GPT-5.4

#585

Post by IgorPartola »

Aurornis wrote: post #387
__jl__ wrote: post #335

What a model mess! OpenAI now has three price points: GPT 5.1, GPT 5.2 and now GPT 5.4. There version numbers jump across different model lines with codex at 5.3, what they now call instant also at 5.3. Anthropic are really the only ones who managed to get this under control: Three models, priced at three different levels. New models are immediately available everywhere. Google essentially only has Preview models! Th…

> What a model mess! OpenAI now has three price points: GPT 5.1, GPT 5.2 and now GPT 5.4. I don't know, this feels unnecessarily nitpicky to me It isn't hard to understand that 5.4 > 5.2 > 5.1. It's not hard to understand that the dash-variants have unique properties that you want to look up before selecting. Especially for a target audience of software engineers skipping a version number is a common occurrence and n…

The issue isn’t 5.4 > 5.2 etc. It is that there is a second dimension which is the model size and a third dimension which is what it is tuned for. And when you are releasing so quickly that flagship your instant mini model is on one numerical version but your flagship tool calling mini model is on another it is confusing trying to figure out which actual model you want for your use case.

It’s not impossible to figure out but it is a symptom of them releasing as quickly as possible to try to dominate the news and mindshare.

Re: GPT-5.4

#586

Post by esafak »

An important feature is the introduction of tool search, which provides models with a "lightweight list of available tools along with a tool search capability", thereby Making MCP Great Again!

Re: GPT-5.4

#587

Post by joegibbs »

creamyhorror wrote: post #223

I've only used 5.4 for 1 prompt (edit: 3@high now) so far (reasoning: extra high, took really long), and it was to analyse my codebase and write an evaluation on a topic. But I found its writing and analysis thoughtful, precise, and surprisingly clearly written, unlike 5.3-Codex. It feels very lucid and uses human phrasing. It might be my AGENTS.md requiring clearer, simpler language, but at least 5.4's doing a good…

The weird phrasing was my biggest gripe with 5.3 so I'm glad they've fixed that up. It couldn't say anything without a heap of impenetrable jargon and it was obsessed with the word "drive". Nothing could cause anything, it had to be "driven".

Re: GPT-5.4

#588

Post by QRe »

XCSme wrote: post #579

Looking ok, but nothing special: https://aibenchy.com/model/openai-gpt-5-4-medium/

Does this LLM benchmark have any actual credibility? I get why they chose to not publish the actual tests but I find it highly dubious that there are only 15 tests and Gemini 3 Flash performs best.

Re: GPT-5.4

#589

Post by sillysaurusx »

FrankBooth wrote: post #305
sillysaurusx wrote: post #251

Earlier quoted context omitted.

You may want to look over this thread from cperciva: https://x.com/cperciva/status/2029645027358495156 I too tried Codex and found it similarly hard to control over long contexts. It ended up coding an app that spit out millions of tiny files which were technically smaller than the original files it was supposed to optimize, except due to there being millions of them, actual hard drive usage was 18x larger. It seemed…

What’s the connection with context size in that thread? It seems more like an instruction following problem.

It occurred to me that searching 196 .c files was a context window issue, but maybe there’s something else going on. Either way, Codex could behave better.

Re: GPT-5.4

#590

Post by gck1 »

nowittyusername wrote: post #330
tedsanders wrote: post #59

Earlier quoted context omitted.

Yeah, long context vs compaction is always an interesting tradeoff. More information isn't always better for LLMs, as each token adds distraction, cost, and latency. There's no single optimum for all use cases. For Codex, we're making 1M context experimentally available, but we're not making it the default experience for everyone, as from our testing we think that shorter context plus compaction works best for most p…

Personally what I am more interested about is effective context window. I find that when using codex 5.2 high, I preferred to start compaction at around 50% of the context window because I noticed degradation at around that point. Though as of a bout a month ago that point is now below that which is great. Anyways, I feel that I will not be using that 1 million context at all in 5.4 but if the effective window is som…

Not sure how accurate this is, but found contextarena benchmarks today when I had the same question.

It appears only gemini has actual context == effective context from these. Although, I wasn't able to test this neither in gemini cli, nor antigravity with my pro subscription because, well, it appears nobody actually uses these tools at Google.

https://contextarena.ai/?showLabels=false

Post reply on HN