Smaller, faster, safer: running Kimi and GLM at scale

Posted by ascorbic 6 hours ago

Counter121Comment31OpenOriginal

Comments

Comment by scrlk 4 hours ago

Nice to see a provider being transparent about KV cache quantisation. I've been suspecting that some providers do this silently whilst heavily promoting their unquantised weights, even though KV quantisation can degrade quality more than weight quantisation.

However, I wish their testing were more detailed. Firstly, some model families are more sensitive to KV quantisation than others (only Kimi K2.6 was tested). Secondly, the evaluation suite they use to claim that FP8 KV quantisation is indistinguishable is noticeably lacking coding benchmarks; in long-running tasks, minor tool call errors compound over time.

Comment by amluto 2 hours ago

They made an extremely strong claim:

> None of this would matter if it changed the model's answers

If they want to assert that the answers don’t change, then perhaps they should calculate the statistical distance between the token probability outputs or something to that effect. I doubt the results would indicate that the answers don’t change by any reasonable interpretation.

Maybe the results are still good enough.

Comment by scrlk 1 hour ago

KL divergence is your friend when it comes to evaluating the effects of quantisation: https://en.wikipedia.org/wiki/Kullback%E2%80%93Leibler_diver...

Comment by anonova 2 hours ago

vLLM's study also concluded that "FP8 can deliver meaningful latency and capacity gains with small or negligible accuracy loss". Their benchmarks include LiveCodeBench 6.

https://vllm-project.github.io/2026/04/22/fp8-kvcache.html

Comment by 55 minutes ago

Comment by syntaxing 3 hours ago

> View pricing in the Cloudflare dashboard ↗

Why… I wanted to see if it’s worth it to use cloudflare’s endpoint but I can’t even see the pricing

Comment by maxmcd 1 hour ago

I'm not sure how accurate this is, but there is pricing here: https://openrouter.ai/provider/cloudflare

Comment by culi 23 minutes ago

So they don't even support K3? What's the point. K2.7 Code is practically free already

Comment by joshuamcginnis 1 hour ago

> If squeezing the best open models onto GPUs and serving them to millions of developers sounds like your kind of problem, come work with us.

What is the typical job title and/or skillset for this?

Comment by simonw 1 hour ago

I've seen this called MLOps.

Comment by brokenodo 4 hours ago

I was interested in reading this until my slop detector went off at the paragraph starting with “It's worth being precise about where the benefit comes from, because it isn't raw speed.”

I love AI, but I really hate reading it.

Comment by hankbond 4 hours ago

I have had to stop commenting this because it would end up on 50% of the posts here. I really wish we could flag prose as ai-generated on here and just filter it out.

Comment by hamdingers 2 hours ago

Drives me nuts that comments are held to a higher standard than submissions.

HN is for conversation between humans[1] (about AI generated blogspam, apparently)

1. https://news.ycombinator.com/newsguidelines.html

Comment by mannanj 1 hour ago

Meta recently added a filter as a requirement for posts on Facebook. if it was ai generated, you are required to check off a box for that on your posts.

I've been asking for that for some time.

Comment by HDBaseT 1 minute ago

This is only to help them filter out AI generations for their own training data. There is no way you can "block" all "ai generated content" from your view.

Comment by dgellow 3 hours ago

Don’t stop commenting about it, if there is something we (the readers) can do is ensure it is seen as uncool to post slop content

Comment by serf 2 hours ago

fine but please realize that also just pushes people away from caring about the comment section at all, too.

I can only read so many (either pro OR against) ".. IT'S AI! ! !" comments before skipping the thread. I can't be the only one.

Comment by gr_norm 3 hours ago

LinkedIn (of all places!) announced a button for flagging this recently: https://www.linkedin.com/posts/hsrinivasan1_ai-slop-is-a-top...

How well it would work on this site, I'm not sure.

Comment by speedgoose 3 hours ago

If it works, it’s going to be the best feature introduced by a social network in a long time. Incredible that it comes from LinkedIn.

Comment by hankbond 3 hours ago

Next up, LinkedIn starts using this feedback to train a classifier. They then announce an officially approved "not slop" classification only for LinkedIn Gold member posts. The classified posts have a wider reach due to everyone filtering out AI slop. Non-members automatically get bucketed in with the slop bc they don't pay to have the verified classifier run on them.

Comment by serf 2 hours ago

seems like at that point it'd be a bigger benefit for LinkedIn to provide the one and only model trained to subvert the human capability to detect slop using that new negative corpus.

then we're all screwed I guess. lol

Comment by physix 3 hours ago

Better would have been to offer a button to flag something that does NOT seem like AI slop on LinkedIn.

Comment by Oras 2 hours ago

If there is an action on AI slop on LI, it will end up with almost no posts at all

Comment by trollbridge 3 hours ago

Sign up for Pangram and install the browser extension; covers X, Reddit, and Substack, and more to come.

Comment by TacticalCoder 1 hour ago

> “It's worth being precise about where the benefit comes from, because it isn't raw speed.“

What's funny it's that is as if AI "learned" to speak english but not really. People simply don't speak using those strange constructs: those sentences sound a bit like if a "Karen" was trying to make a point.

What's scary, to me, as a dev using AI, is that those LLMs do the same thing with code: it looks like proper code, but it really ain't so once you dig a bit.

It's verbose and doesn't add anything: it's just infinite verbiage / sloppy-pasta.

Crazy thing though it's that it's 2026 and apparently devs can't be bothered to copy/paste their sloppy-pasta LLMish into a de-sloppifier before publishing blog posts.

Comment by mrkstu 2 hours ago

I'm getting somewhat conflicted. There is a Twitter account I follow that has great takes and lots of deeply personal posts... BUT he obviously uses an LLM in his writing pipeline somewhere. Too many AIisms scattered about to ignore.

I'm not thrilled with it, but he is obviously using it to improve his writing overall- to communicate some great ideas that are personal and germane. I've decided that being too inflexible serves no one. If it is true slop, I'll not revisit the writer in the future- if they are using AI to polish writing that at its core is a unique voice, I'll accept it and learn to live with it...

Comment by colingauvin 2 hours ago

Came here to comment this.

Comment by arjie 3 hours ago

Cloudflare blogs are not meant to be human-read, AFAIK. They're raw material meant to be fed into an agent to be filtered down. I rarely read the contents because they are usually word-expanded to a greater degree than an article from The Atlantic.

Comment by dgellow 3 hours ago

That’s disappointing, in the past cloudflare had some of the best engineering blog articles

Comment by arjie 2 hours ago

I don't disagree, but at some point in the last year they ended up severely word-expanded. So in a revealed sense, they are no longer meant for human consumption except for those who don't significantly value their own time. There is very little information in the post that an agent can't pull for you:

* they use quantized models

* they quantize KV cache

* they have a cache tagging mechanism to prevent cache misuse (neat)

The agent can extract numbers without filler prose as well.

Comment by lostmsu 2 hours ago

So they quantize models, only tell about it in the blog post (instead of a warning on the model page), and even in the blog post pretend there's no difference by benchmarking on small context tasks many of which are saturated. Coding agents will probably be severely negatively affected by KV quantization.

I'd say serving quantized models without saying so on the "store" page is fraud.

Comment by whimsicalism 1 hour ago

don't disagree, but there is a big difference between 'quantized model/weights' and quantized activations