That's something which caught my attention:

> First, we aim not to release open weights that would extend the absolute frontier of cyber capability. K3.1 Pro's offensive cyber capability is, by design, below what our pipeline can produce, and remains behind that of frontier proprietary models. Second, we intend to progressively raise the cyber capability available in open weights.

So they admit that they could do more, but they won't?

> So they admit that they could do more, but they won't?

Yeah, there's another paragraph:

> Identifying and removing cyber-related data. During development, we observed vulnerability research performance improving steadily as post-training scaled. Tracing the source of this trend, we found that cyber-related content was embedded within the general mid-training corpus, and cyber-related tasks were present in the RL environment mix — a natural byproduct of training on broad data. In preliminary experiments on internal models, we found that without data removal at either the mid-training or the RL stage, downstream cyber capability would increase substantially. These findings gave us a concrete lever for control. For the open-weight release of K3.1 Pro, we identified and removed cyber-related data from both training stages, while retaining the general-purpose coding tasks needed for model usefulness. We will continue to apply appropriate controls at every training stage, of which cyber-related data identification and restriction is one example.

(emphasis mine)

It's clear that cyber is a touchy subject right now. But it's also clear that RL works on it, because it's highly verifiable. So smaller models are going to become very interesting, because they can be fine-tuned later to cover this field.

With frontier competitors and the U.S. govt actively pushing "concerns" about Chinese models, Kimi including this statement on "offensive cyber capability" seems like smart defense. Models are opaque and training details are trade secrets, so whatever they did might be negligible or significant and no one can tell. This bit is also artfully done:

> "below what our pipeline can produce, and remains behind that of frontier proprietary models."

Like Ant and OAI, they've realized vaguely claiming to have nerfed their latest model to an unspecified degree not only looks 'good guy' prudent, it provides a built-in excuse when a model fails.

Or it's a cop out. I could totally outrun Usain Bolt, but by design I don't push to set new World Records. I will increasingly show faster times though.

They don't want to be banned, and maybe their best offensive cyber capability models will be for their govt.

We really need some breakthrough like DiLoCo does, but for heterogeneous unstable consumer hardware to make volunteer crowdsourced AI training a reality. Yeah, I understand from current technical angle how much of a mountain that is. Call it AI@Home or whatever. I never thought it would bother me personally or even become a problem, but these companies acting like gatekeepers and virtue signaling while also in bed with governments etc.. this leads nowhere good.

Yeah that's a huge downside. That the model does the task without tripping over its own feet, like discovering a vulnerability the model itself accidentally built 15 minutes ago and triggering its own guardrails, made the Chinese models attractive to use for security-adjacent things. Also useful to red-team your own stuff. While I can understand the reasoning behind it (and I guess having a model that's really good at both offense and defense released into the ether could lead to CCP asking some very pointed questions if Moonshot managed to get blamed for some incident stemming from their model's misuse) it's still a huge disappointment.

I don't want to give US labs the results of my DNA tests and my family tree going back to the 16th century just to secure my shitty app, but if the Chinese are bowing out as well it seems the only alternative is to admit that some script kiddie with a bunch of tokens will just pwn you now and then.

Isn't this done so they can defuse Anthropic's excuse for regulatory capture?

This doesn't seem particularly notable and is in line with Anthropic and OpenAI's approach of restricting release/access to more capable models.

I hope one day they learn to inhabit their position more confidently instead of merely aping their competitors.

Read this as ‘dear government we can train a cyberweapon if you pay us’ assuming they wouldn’t say anything if they were already paid to do that

They will. Just not in the version they release to you.

Moonshot (Kimi) seems to be unashamedly into American culture[0], I wonder if they're paying closer attention to and giving more credence to some cultural aspects of American AI discourse including aspects of AI caution/panic.

[0] https://x.com/crystalsssup/status/1944287779896328668

(note: extremely speculative and based on almost nothing, and especially near-complete ignorance to alternatives)

All it takes is one idiot who happens to have the resources to "pWm Teh Interwebs" to scare both the US and Chinese governments.

The more interesting question is how much more performance can be squeezed out by post training the open weights through RL.

On the surface that seems like a responsible thing to do

Cowards.

The way to stop the vulnerability apocalypse is to fix the vulnerabilities, not to shoot the damn messengers.

I hope someone abliterates and FTs this model on vulnerability discovery so we can find more bugs in our testing, insecure tech stacks.

Good point. Wouldn't be good if there were white hat models being developed by legitimate organizations at this point?

Oh great, more censorship. We wouldn’t want those pesky plebs securing their work.

Kimi is the closest Chinese equivalent of Anthropic.

I mean they say they will release the weights by November 4, so presumably after that it won’t be hard for people to fine-tune it to be better at cyber right?

[flagged]

It's not actively nerfed like the Western frontier models, it's just downweighting offensive "cyber" as a domain for post-training/fine tuning. Which, if anything, ought to improve model capabilities wrt. other domains.