OpenAI O3-Mini • Hacker News Board

Re: OpenAI O3-Mini

#131

Post by observationist » Fri, Jan 31, 2025, 7:39 PM UTC

ryanhecht wrote: post #39

> While OpenAI o1 remains our broader general knowledge reasoning model, OpenAI o3-mini provides a specialized alternative for technical domains requiring precision and speed. I feel like this naming scheme is growing a little tired. o1 is for general knowledge reasoning, o3-mini replaces o1-mini but might be more specialized than o1 for certain technical domains...the "o" in "4o" is for "omni" (referring to its mult…

They should be calling it ChatGPT and ChatGPT-mini, with other models hidden behind some sort of advanced mode power user menu. They can roll out major and minor updates by number. The whole point of differentiating between models is to get users to self limit the compute they consume - rate limits make people avoid using the more powerful models, and if they have a bad experience using the less capable models, or if they're frustrated by hopping between versions without some sort of nuanced technical understanding, it's just a bad experience overall.

OpenAI is so scattered they haven't even bothered using their own state of the art AI to come up with a coherent naming convention? C'mon, get your shit together.

Re: OpenAI O3-Mini

#132

Post by not_a_bot_4sho » Fri, Jan 31, 2025, 7:39 PM UTC

siliconc0w wrote: post #42

The real heated contest here amongst the top AI labs is to see who can come up with the most confusing product names.

Someone dropped the ball with Phi models. There is clearly an opportunity for XP and Ultimate and X/S editions.

Re: OpenAI O3-Mini

#133

Post by koakuma-chan » Fri, Jan 31, 2025, 7:39 PM UTC

ninetyninenine wrote: post #110

koakuma-chan wrote: post #101

Earlier quoted context omitted.

Name is just a label. It's not supposed to mean anything.

Think how awesome the world would be if labels ALSO had meanings.

As someone else said in another thread, if you could derive the definition from a word, the word would be as long as the definition, which would defeat the purpose.

Re: OpenAI O3-Mini

#134

Post by cbg0 » Fri, Jan 31, 2025, 7:40 PM UTC

leovander wrote: post #113

rafaquintanilha wrote: post #104

Earlier quoted context omitted.

You know you are running an extremely nerfed version of the model, right?

I did update my comment, but said that I am using the distilled version, so yes?

Even the full model scores below Claude on livebench so a distilled version will likely be even worse.

Re: OpenAI O3-Mini

#135

Post by reissbaker » Fri, Jan 31, 2025, 7:40 PM UTC

xyzzy9563 wrote: post #16

What is the comparison of this versus DeepSeek in terms of good results and cost?

Probably a good idea to wait for external benchmarks like Aider, but my guess is it'll be somewhere between DeepSeek V3 and R1 in terms of benchmarks — R1 trades blows with o1-high, and V3 is somewhat lower — but I'd expect o3-mini to be considerably faster. Despite the blog post saying paid users can access o3-mini today, I don't see it as an option yet in their UI... But IIRC when they announced o3-mini in December they claimed it would be similar to 4o in terms of overall latency, and 4o is much faster than V3/R1 currently.

Re: OpenAI O3-Mini

#136

Post by throwaway314155 » Fri, Jan 31, 2025, 7:40 PM UTC

turnsout wrote: post #30

Hmm, not seeing it in my dashboard yet (Tier 4)

This has happened to me with (I think) every single major model release (llm or image gen) from OpenAI. They just lie in their release announcements which leaves people scrambling on the day of.

Re: OpenAI O3-Mini

#137

Post by meiraleal » Fri, Jan 31, 2025, 7:40 PM UTC

golly_ned wrote: post #82

kingnothing wrote: post #55

Earlier quoted context omitted.

They really need someone in marketing. If the model is for technical stuff, then call it the technical model. How is anyone supposed to know what these model names mean? The only page of theirs attempting to explain this is a total disaster. https://platform.openai.com/docs/models

Yes, this $300Bn company generating +$3.4Bn in revenue needs to hire marketing expert. They can begin by sourcing ideas from us here to save their struggling business from total marketing disaster.

[dead]

Re: OpenAI O3-Mini

#138

Post by throwaway314155 » Fri, Jan 31, 2025, 7:40 PM UTC

dang wrote: post #18

vincentpants wrote: post #5

Wow, it got to the top of the front page so fast! Weird!

I took a quick look at the data and FWIW the votes look legit to me, if that's what you were wondering.

I'm fairly certain it was sarcasm.

Re: OpenAI O3-Mini

#139

Post by Rooster61 » Fri, Jan 31, 2025, 7:41 PM UTC

xinayder wrote: post #97

"oh no DeepSeek copied our product it's not fair" > proceeds to release a product based on DeepSeek ah, alas the hypocrisy...

The thing they previewed back in December before the whole Deepseek kerfuffle this week?

Don't get me wrong, I'm laughing at OpenAI just like everyone else, but if they were really copying Deepseek, they'd be releasing a smaller model distilled from Deepseek API responses, and have it be open source to boot. This is neither

Re: OpenAI O3-Mini

#140

Post by reissbaker » Fri, Jan 31, 2025, 7:41 PM UTC

evertedsphere wrote: post #22

>developer messages looks like finally their threat model has been updated to take into account that the user might be too "unaligned" to be trusted with the ability to provide a system message of their own

...I'm pretty sure they just renamed the key...