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

GPT-4.1 in the API

openai.com

Post reply on HN
Print view View original

331–340 of 513 posts

Re: GPT-4.1 in the API

#331

Post by resters »

motoboi wrote: post #327
resters wrote: post #325

Earlier quoted context omitted.

I use them as follows: o1-pro: anything important involving accuracy or reasoning. Does the best at accomplishing things correctly in one go even with lots of context. deepseek R1: anything where I want high quality non-academic prose or poetry. Hands down the best model for these. Also very solid for fast and interesting analytical takes. I love bouncing ideas around with R1 and Grok-3 bc of their fast responses and…

You probably know this but it can already generate accurate diagrams. Just ask for the output in a diagram language like mermaid or graphviz

I've had mixed and inconsistent results and it hasn't been able to iterate effectively when it gets close. Could be that I need to refine my approach to prompting. I've tried mermaid and SVG mostly, but will also try graphviz based on your suggestion.

Re: GPT-4.1 in the API

#332

Post by aoeusnth1 »

swyx wrote: post #226
simonw wrote: post #211

Earlier quoted context omitted.

I'm surprised and a little disappointed by the result concerning instructions at the top, because it's incompatible with prompt caching: I would much rather cache the part of the prompt that includes the long document and then swap out the user question at the end.

yep. we address it in the podcast. presumably this is just a recent discovery and can be post-trained away.

If you're skimming a text to answer a specific question, you can go a lot faster than if you have to memorize the text well enough to answer an unknown question after the fact.

Re: GPT-4.1 in the API

#334

Post by resters »

bangaladore wrote: post #329
motoboi wrote: post #327

Earlier quoted context omitted.

You probably know this but it can already generate accurate diagrams. Just ask for the output in a diagram language like mermaid or graphviz

My experience is it often produces terrible diagrams. Things clearly overlap, lines make no sense. I'm not surprised as if you told me to layout a diagram in XML/YAML there would be obvious mistakes and layout issues. I'm not really certain a text output model can ever do well here.

FWIW I think a multimodal model could be trained to do extremely well with it given sufficient training data. A combination of textual description of the system and/or diagram, source code (mermaid, SVG, etc.) for the diagram, and the resulting image, with training to translate between all three.

Re: GPT-4.1 in the API

#335

Post by fragmede »

sottol wrote: post #277
observationist wrote: post #265

Earlier quoted context omitted.

Anyone making claims with a horizon beyond two months about structure or capabilities will be wrong - it's sama's job to show confidence and vision and calm stakeholders, but if you're paying attention to the field, the release and research cycles are still contracting, with no sense of slowing any time soon. I've followed AI research daily since GPT-2, the momentum is incredible, and even if the industry sticks with…

Maybe that's true for absolute arm-chair-engineering outsiders (like me) but these models are in training for months, training data is probably being prepared year(s) in advance. These models have a knowledge cut-off in 2024 - so they have been in training for a while. There's no way sama did not have a good idea that this non-COT was in the pipeline 2 months ago. It was probably finished training then and undergoing…

With new hardware from Nvidia announced coming out, those months turn into weeks.

Re: GPT-4.1 in the API

#336

Post by omneity »

christianqchung wrote: post #260
omneity wrote: post #245

Earlier quoted context omitted.

I'm not sure to which extent this opinion is accurately informed. It is well known that nobody trains on 1M token-long content. It wouldn't work anyway as the dependencies are too far fetched and you end up with vanishing gradients. RoPE (Rotary Positional Embeddings, think modulo or periodic arithmetics) scaling is key, whereby the model is trained on 16k tokens long content, and then scaled up to 100k+ [0]. Qwen 1M…

But Llama 4 Scout does badly on long context benchmarks despite claiming 10M. It scores 1 slot above Llama 3.1 8B in this one[1]. [1] https://github.com/adobe-research/NoLiMa

Indeed, but it does not take away the fact that long context is not trained through long content but by scaling short content instead.

Re: GPT-4.1 in the API

#337

Post by throwup238 »

lxgr wrote: post #317

As a ChatGPT user, I'm weirdly happy that it's not available there yet. I already have to make a conscious choice between - 4o (can search the web, use Canvas, evaluate Python server-side, generate images, but has no chain of thought) - o3-mini (web search, CoT, canvas, but no image generation) - o1 (CoT, maybe better than o3, but no canvas or web search and also no images) - Deep Research (very powerful, but I have…

> - Deep Research (very powerful, but I have only 10 attempts per month, so I end up using roughly zero)

Same here, which is a real shame. I've switched to DeepResearch with Gemini 2.5 Pro over the last few days where paid users have a 20/day limit instead of 10/month and it's been great, especially since now Gemini seems to browse 10x more pages than OpenAI Deep Research (on the order of 200-400 pages versus 20-40).

The reports are too verbose but having it research random development ideas, or how to do something particularly complex with a specific library, or different approaches or architectures to a problem has been very productive without sliding into vibe coding territory.

Re: GPT-4.1 in the API

#338

Post by XCSme »

XCSme wrote: post #274

I tried 4.1-mini and 4.1-nano. The response are a lot faster, but for my use-case they seem to be a lot worse than 4o-mini(they fail to complete the task when 4o-mini could do it). Maybe I have to update my prompts...

Even after updating my prompts, 4o-mini still seems to do better than 4.1-mini or 4.1-nano for a data-processing task.

Re: GPT-4.1 in the API

#339

Post by omneity »

kmeisthax wrote: post #292
omneity wrote: post #245

Earlier quoted context omitted.

I'm not sure to which extent this opinion is accurately informed. It is well known that nobody trains on 1M token-long content. It wouldn't work anyway as the dependencies are too far fetched and you end up with vanishing gradients. RoPE (Rotary Positional Embeddings, think modulo or periodic arithmetics) scaling is key, whereby the model is trained on 16k tokens long content, and then scaled up to 100k+ [0]. Qwen 1M…

Is there any evidence that GPT-4.1 is using RoPE to scale context? Also, I don't know about Qwen, but I know Llama 4 has severe performance issues, so I wouldn't use that as an example.

I am not sure about public evidence. But the memory requirements alone to train on 1M long windows would make it a very unrealistic proposition compared to RoPE scaling. And as I mentioned RoPE is essential for long context anyway. You can't train it in the "normal way". Please see the paper I linked previously for more context (pun not intended) on RoPE.

Re: Llama 4, please see the sibling comment.

Re: GPT-4.1 in the API

#340

Post by shortcord »

resters wrote: post #325
lxgr wrote: post #317

As a ChatGPT user, I'm weirdly happy that it's not available there yet. I already have to make a conscious choice between - 4o (can search the web, use Canvas, evaluate Python server-side, generate images, but has no chain of thought) - o3-mini (web search, CoT, canvas, but no image generation) - o1 (CoT, maybe better than o3, but no canvas or web search and also no images) - Deep Research (very powerful, but I have…

I use them as follows: o1-pro: anything important involving accuracy or reasoning. Does the best at accomplishing things correctly in one go even with lots of context. deepseek R1: anything where I want high quality non-academic prose or poetry. Hands down the best model for these. Also very solid for fast and interesting analytical takes. I love bouncing ideas around with R1 and Grok-3 bc of their fast responses and…

Gemini 2.5 Pro is quite good at code.

Has become my go to for use in Cursor. Claude 3.7 needs to be restrained too much.

Post reply on HN