Kimi K3: Open Frontier Intelligence

Posted by vincent_s 15 hours ago

Kimi K3: Open Frontier Intelligence(www.kimi.com)

https://www.kimi.com/en

Kimi K3 Intelligence, Performance & Price Analysis: https://artificialanalysis.ai/models/kimi-k3

1362 points | 835 commentspage 9

ls612 4 hours ago|

https://www.reuters.com/world/china/chinas-xi-outline-ai-dip...

Reuters is reporting that Xi is planning to endorse open source/weights AI in a speech tomorrow. This is probably highly relevant to why Moonshot is committing to making Kimi open weights.

jdw64 9 hours ago||

What subscription plan for Kimi 3 would be the most cost effective? Most people only talk about API efficiency, but is there any place that evaluates how much you get with the subscription plans?

dwa3592 8 hours ago||

waiting for - "Running Kimi K3 on X years old hardware".

luciana1u 10 hours ago||

at this rate the next model release will just be a git commit hash and a shrug emoji

matheusmoreira 5 hours ago||

Holy fuck. This is an open weights model? It's literally trailing the frontier models! This is insane!

It's my dream to own hardware that can run this!

sudosysgen 10 hours ago||

https://www.kimi.com/blog/kimi-k3

"The full model weights will be released by July 27, 2026."

benjiro29 11 hours ago||

Full benchmarks in Mandarin:

https://mp.weixin.qq.com/s/V4xhEIy8xDXSMDPrPkmUAQ

Translation:

https://mp-weixin-qq-com.translate.goog/s/V4xhEIy8xDXSMDPrPk...

Cheaper then GPT 5.6 Sol (according to their results) ...

aussieguy1234 6 hours ago||

I'll switch to this for now.

I'm expecting Anthropics reply soon though.

It would be trivial for them to distill Mythos.

ed_mercer 6 hours ago|

AFAIKI Fable = Mythos but with more guardrails

vblanco 10 hours ago|

Another deepseek moment? it seems they have fully caught with fable tier of models, and this was a lot sooner than was expected.

InsideOutSanta 10 hours ago|

Yeah, I would have expected Zhipu to ship a Fable-adjacent model by the end of the year, but the jump from Kimi 2.7 (which I think is just barely at the level where it is genuinely helpful for coding) to this is absolutely bonkers. And this is clearly not just benchmaxing; this thing actually works.

If you told me I could only use this and never use Fable or Sol again, I'd shrug and not feel like I'd lost much.

procgen 10 hours ago|||

Now it seems the best way to tell if a frontier model is benchmaxxed is to check if it can autonomously solve a major open mathematical problem.

re-thc 10 hours ago|||

> Yeah, I would have expected Zhipu to ship a Fable-adjacent model by the end of the year

There were talks of a GLM 5.3 in August, so maybe not that far away...

More comments...