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

Open models by OpenAI

openai.com

Post reply on HN
Print view View original

51–60 of 909 posts

Re: Open models by OpenAI

#51

Post by Quarrel »

Disposal8433 wrote: post #21

Please don't use the open-source term unless you ship the TBs of data downloaded from Anna's Archive that are required do build it yourself. And dont forget all the system prompts to censor the multiple topics that they don't want you to see.

Is your point really that- "I need to see all data downloaded to make this model, before I can know it is open"? Do you have $XXB worth of GPU time to ingest that data with a state of the art framework to make a model? I don't. Even if I did, I'm not sure FB or Google are in any better position to claim this model is or isn't open beyond the fact that the weights are there.

They're giving you a free model. You can evaluate it. You can sue them. But the weights are there. If you dislike the way they license the weights, because the license isn't open enough, then sure, speak up, but because you can't see all the training data??! Wtf.

Re: Open models by OpenAI

#54

Post by BoorishBears »

johntiger1 wrote: post #10

Wow, this will eat Meta's lunch

Maverick and Scout were not great, even with post-training in my experience, and then several Chinese models at multiple sizes made them kind of irrelevant (dots, Qwen, MiniMax)

If anything this helps Meta: another model to inspect/learn from/tweak etc. generally helps anyone making models

Re: Open models by OpenAI

#55

Post by petuman »

artembugara wrote: post #14

Disclamer: probably dumb questions so, the 20b model. Can someone explain to me what I would need to do in terms of resources (GPU, I assume) if I want to run 20 concurrent processes, assuming I need 1k tokens/second throughput (on each, so 20 x 1k) Also, is this model better/comparable for information extraction compared to gpt-4.1-nano, and would it be cheaper to host myself 20b?

> assuming I need 1k tokens/second throughput (on each, so 20 x 1k)

3.6B activated at Q8 x 1000 t/s = 3.6TB/s just for activated model weights (there's also context). So pretty much straight to B200 and alike. 1000 t/s per user/agent is way too fast, make it 300 t/s and you could get away with 5090/RTX PRO 6000.

Re: Open models by OpenAI

#56

Post by 4b6442477b1280b »

minimaxir wrote: post #20

I'm disappointed that the smallest model size is 21B parameters, which strongly restricts how it can be run on personal hardware. Most competitors have released a 3B/7B model for that purpose. For self-hosting, it's smart that they targeted a 16GB VRAM config for it since that's the size of the most cost-effective server GPUs, but I suspect "native MXFP4 quantization" has quality caveats.

with quantization, 20B fits effortlessly in 24GB

with quantization + CPU offloading, non-thinking models run kind of fine (at about 2-5 tokens per second) even with 8 GB of VRAM

sure, it would be great if we could have models in all sizes imaginable (7/13/24/32/70/100+/1000+), but 20B and 120B are great.

Re: Open models by OpenAI

#57

Post by thegeomaster »

amarcheschi wrote: post #39
Leary wrote: post #28

GPQA Diamond: gpt-oss-120b: 80.1%, Qwen3-235B-A22B-Thinking-2507: 81.1% Humanity’s Last Exam: gpt-oss-120b (tools): 19.0%, gpt-oss-120b (no tools): 14.9%, Qwen3-235B-A22B-Thinking-2507: 18.2%

Glm 4.5 seems on par as well

GLM-4.5 seems to outperform it on TauBench, too. And it's suspicious OAI is not sharing numbers for quite a few useful benchmarks (nothing related to coding, for example).

One positive thing I see is the number of parameters and size --- it will provide more economical inference than current open source SOTA.

Re: Open models by OpenAI

#58

Post by layer8 »

outlore wrote: post #37
Disposal8433 wrote: post #21

Please don't use the open-source term unless you ship the TBs of data downloaded from Anna's Archive that are required do build it yourself. And dont forget all the system prompts to censor the multiple topics that they don't want you to see.

by your definition most of the current open weight models would not qualify

That’s why they are called open weight and not open source.

Re: Open models by OpenAI

#59

Post by mlyle »

artembugara wrote: post #14

Disclamer: probably dumb questions so, the 20b model. Can someone explain to me what I would need to do in terms of resources (GPU, I assume) if I want to run 20 concurrent processes, assuming I need 1k tokens/second throughput (on each, so 20 x 1k) Also, is this model better/comparable for information extraction compared to gpt-4.1-nano, and would it be cheaper to host myself 20b?

An A100 is probably 2-4k tokens/second on a 20B model with batched inference.

Multiply the number of A100's you need as necessary.

Here, you don't really need the ram. If you could accept fewer tokens/second, you could do it much cheaper with consumer graphics cards.

Even with A100, the sweet-spot in batching is not going to give you 1k/process/second. Of course, you could go up to H100...

Re: Open models by OpenAI

#60

Post by tedivm »

x187463 wrote: post #23

Running a model comparable to o3 on a 24GB Mac Mini is absolutely wild. Seems like yesterday the idea of running frontier (at the time) models locally or on a mobile device was 5+ years out. At this rate, we'll be running such models in the next phone cycle.

It only seems like that if you haven't been following other open source efforts. Models like Qwen perform ridiculously well and do so on very restricted hardware. I'm looking forward to seeing benchmarks to see how these new open source models compare.
Post reply on HN