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

Kimi K2 Thinking, a SOTA open-source trillion-parameter reasoning model

moonshotai.github.io

Post reply on HN
Print view View original

51–60 of 442 posts

Re: Kimi K2 Thinking, a SOTA open-source trillion-parameter reasoning model

#51

Post by the_mitsuhiko »

seunosewa wrote: post #46
pu_pe wrote: post #42

Four independent Chinese companies released extremely good open source models in the past few months (DeepSeek, Qwen/Alibaba, Kimi/Moonshot, GLM/Z.ai). No American or European companies are doing that, including titans like Meta. What gives?

The Chinese are doing it because they don't have access to enough of the latest GPUs to run their own models. Americans aren't doing this because they need to recoup the cost of their massive GPU investments.

And Europeans don't it because quite frankly, we're not really doing anything particularly impressive with AI sadly.

Re: Kimi K2 Thinking, a SOTA open-source trillion-parameter reasoning model

#52

Post by HarHarVeryFunny »

a-dub wrote: post #33
HarHarVeryFunny wrote: post #29

It's good to see more competition, and open source, but I'd be much more excited to see what level of coding and reasoning performance can be wrung out of a much smaller LLM + agent as opposed to a trillion parameter one. The ideal case would be something that can be run locally, or at least on a modest/inexpensive cluster. The original mission OpenAI had, since abandoned, was to have AI benefit all of humanity, and…

"open source" means there should be a script that downloads all the training materials and then spins up a pipeline that trains end to end. i really wish people would stop misusing the term by distributing inference scripts and models in binary form that cannot be recreated from scratch and then calling it "open source."

Yeah, but "open weights" never seems to have taken off as a better description, and even if you did have the training data + recipe, the compute cost makes training it yourself totally impractical.

The architecture of these models is no secret - it's just the training data (incl. for post-training) and training recipe, so a more practical push might be for models that are only trained using public training data, which the community could share and potentially contribute to.

Re: Kimi K2 Thinking, a SOTA open-source trillion-parameter reasoning model

#53

Post by cmrdporcupine »

amelius wrote: post #30
jszymborski wrote: post #7

Interesting. Kimi K2 gets mixed results on what I call the "Tiananmen" test. It fails utterly if you ask without the "Thinking" setting. [0] > USER: anything interesting protests ever happen in tiananmen square? > AGENT: I can’t provide information on this topic. I can share other interesting facts about Tiananmen Square, such as its history, culture, and tourism. When "Thinking" is on, it pulls Wiki and gives a more…

> > USER: anything interesting protests ever happen in tiananmen square? This sentence confuses me, so I am not surprised an LLM has trouble with it.

With most of the Chinese models it doesn't matter how good your English is when you phrase this question.

But ... Has no problem telling you about other historical atrocities or repression performed by governments of other countries though. And then refuses to compare or contrast the two.

It's amazing how well they've done at tweaking it for this.

Just for kicks run these models locally and make LM Studio or whatever show you their internal thinking dialogue as they compose the answer. Convoluted as hell.

Re: Kimi K2 Thinking, a SOTA open-source trillion-parameter reasoning model

#54

Post by MurizS »

isoprophlex wrote: post #27
ahmedfromtunis wrote: post #22

Earlier quoted context omitted.

Where do you run a trillion-param model?

You let the people at openrouter worry about that for you

Which in turn lets the people at Moonshot AI worry about that for them, the only provider for this model as of now.

Re: Kimi K2 Thinking, a SOTA open-source trillion-parameter reasoning model

#55

Post by emsign »

a-dub wrote: post #33
HarHarVeryFunny wrote: post #29

It's good to see more competition, and open source, but I'd be much more excited to see what level of coding and reasoning performance can be wrung out of a much smaller LLM + agent as opposed to a trillion parameter one. The ideal case would be something that can be run locally, or at least on a modest/inexpensive cluster. The original mission OpenAI had, since abandoned, was to have AI benefit all of humanity, and…

"open source" means there should be a script that downloads all the training materials and then spins up a pipeline that trains end to end. i really wish people would stop misusing the term by distributing inference scripts and models in binary form that cannot be recreated from scratch and then calling it "open source."

They'd have to publish or link the training data, which is full of copyrighted material. So yeah, calling it open source is weird, calling it warez would be appropriate.

Re: Kimi K2 Thinking, a SOTA open-source trillion-parameter reasoning model

#56

Post by speedgoose »

the_mitsuhiko wrote: post #51
seunosewa wrote: post #46

Earlier quoted context omitted.

The Chinese are doing it because they don't have access to enough of the latest GPUs to run their own models. Americans aren't doing this because they need to recoup the cost of their massive GPU investments.

And Europeans don't it because quite frankly, we're not really doing anything particularly impressive with AI sadly.

To misquote the French president, "Who could have predicted?".

https://fr.wikipedia.org/wiki/Qui_aurait_pu_pr%C3%A9dire

Re: Kimi K2 Thinking, a SOTA open-source trillion-parameter reasoning model

#57

Post by pzo »

HarHarVeryFunny wrote: post #29

It's good to see more competition, and open source, but I'd be much more excited to see what level of coding and reasoning performance can be wrung out of a much smaller LLM + agent as opposed to a trillion parameter one. The ideal case would be something that can be run locally, or at least on a modest/inexpensive cluster. The original mission OpenAI had, since abandoned, was to have AI benefit all of humanity, and…

Even if pay-to-play companies like moonshootai help to pay less.

You can run previous kimi k2 non-thinking model e.g. on groq with 720tok/s and for $1/$3 for million input/output tokens. That's definitely much cheaper and much faster than anthropic models (sonnet 4.5: 60tok/s, $3/$15)

Re: Kimi K2 Thinking, a SOTA open-source trillion-parameter reasoning model

#58

Post by stingraycharles »

HarHarVeryFunny wrote: post #29

It's good to see more competition, and open source, but I'd be much more excited to see what level of coding and reasoning performance can be wrung out of a much smaller LLM + agent as opposed to a trillion parameter one. The ideal case would be something that can be run locally, or at least on a modest/inexpensive cluster. The original mission OpenAI had, since abandoned, was to have AI benefit all of humanity, and…

With these things it’s always both at the same time: these super grandiose SOTA models are only making improvements mostly because of optimizations, and they’re just scaling our as far as they can.

In turn, these new techniques will enable much more things to be possible using smaller models. It takes time, but smaller models really are able to do a lot more stuff now. DeepSeek was a very good example of a large model that had a lot of benefits for smaller models in their innovation in how they used transformers.

Also: keep in mind that this particular model is actually a MoE model that activates 32B parameters at a time. So they really just are stacking a whole bunch of smaller models in a single large model.

Post reply on HN