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

GLM-5.3-Flash

z.ai

Post reply on HN
Print view View original

421–430 of 605 posts

Re: GLM-5.3-Flash

#421

Post by efficax »

WarmWash wrote: post #90
sunbum wrote: post #5

> with all of this traffic served on Chinese AI chips RIP Nivida shareholders

I don't see a situation where subscription payers move outside American LLMs (chatgpt, claude, gemini) And I don't see a situation where serious API payers are OK with handing the Chinese state all their data. Like manufactures of decades past did and learned a hard, even existential, lesson for it. The state mantra has been "Collect and Copy" for a long time now, tech just hasn't had that moment to experience it yet…

you don't have to hand them your data, the models are available so you can run them on bedrock yourself (or use another US housed inference service). and for what it's worth in my job i have access to data that gives a picture of the way companies are doing inference, and they're using a lot of chinese models (deepseek-v4 is a huge percentage of inference requests for example)

Re: GLM-5.3-Flash

#422

Post by matheusmoreira »

throw10920 wrote: post #410
deadbabe wrote: post #239

Earlier quoted context omitted.

> Vague prohibitions on "disturbing" or "inappropriate" content, whatever that is. I have prompted out a lot of disturbing and inappropriate content with GLM-5.2, that would have left other American models blanched in the face or clutch their pearls. I think this is mostly a reference to Anti-CCP stuff. In fact, I don't think I've ever even had a prompt refused.

> In fact, I don't think I've ever even had a prompt refused. I very much have. I've gotten GLM-5.2 refusals for extremely benign security testing on my own infrastructure of the same flavor that people were getting (wrongly) flagged for on Fable during the initial release.

That's alarming. I want to use these models to red team my own computers. How are people getting around this?

Re: GLM-5.3-Flash

#423

Post by throw10920 »

matheusmoreira wrote: post #422
throw10920 wrote: post #410

Earlier quoted context omitted.

> In fact, I don't think I've ever even had a prompt refused. I very much have. I've gotten GLM-5.2 refusals for extremely benign security testing on my own infrastructure of the same flavor that people were getting (wrongly) flagged for on Fable during the initial release.

That's alarming. I want to use these models to red team my own computers. How are people getting around this?

> I want to use these models to red team my own computers.

Exactly what I was trying to use it for! ):

I'm in the same boat - I haven't heard of a way to get around it aside from either self-hosting (GLM-5.2? good luck) or "self-hosting" (paying bucks per hour to Vast) an abliterated model.

Re: GLM-5.3-Flash

#424

Post by martinwright »

tyre wrote: post #417
nylonstrung wrote: post #340

Earlier quoted context omitted.

The next 12 months will see OAI and Anthropic spiral into into increasingly hyperbolic PR stunts, manufactured benchmarks and underhanded attempts at regulatory captures I'm sure they have nothing to rival this on a price/performance basis and have already given up on that

> I'm sure they have nothing to rival this on a price/performance basis and have already given up on that How can you be sure about this? They have unbelievable capital. OpenAI is starting to preview its own chips, which could dramatically change the price/performance. We don't know what else Anthropic has cooked up right now that could rival this if they wanted to. Yes, others will _also_ continue to innovate, but m…

w.r.t. the OAI chips, wouldn't they be subject to the same bottlenecks that has plagued semis lately or at least be forced to pay a pretty premium to circumvent that?

Re: GLM-5.3-Flash

#425

Post by guybedo »

minraws wrote: post #415
guybedo wrote: post #391

although i initially thought it didn't make sense financially to run this kind of model locally, i did run the numbers and for heavy users this could justify buying $10k worth of hardware with a ROI over a few months, less than a year. I was looking at my token usage, mostly from subsidized codex/grok subscriptions and i'm a somewhat heavy user. The thing is i would actually use even more tokens if it wasn't for the…

You aren't going to get nearly as much token usage locally from DGX Sparks or even M5 Ultra (though it might be close, unsure would need to get my mittens on it to clarify). You will get around 2-4 concurrent streams of aggregate tokens at best for such a model and around 0.5B output tokens per month assuming you use loops and run it when you are sleeping. That's 500 (per mill) * 0.5$ = 250$ only at most. Then there…

yeah i mostly agree, especially compared to subsidized subscription cost.

But for a heavy user who has enough work to be done so that the box runs almost 24/7 at say 50tok/sec, the math gets interesting against API prices.

And it can be interesting compared to subscription in the sense that you don't have the quota anymore. That means there's probably a lot of things you're not doing because of the quotas that you could do now.

It depends heavily on the tok/sec obviously and the very best solution financially remains subscriptions. But the idea remains entertaining and not that disconnected from reality

Re: GLM-5.3-Flash

#426

Post by cdurth »

skohan wrote: post #228
Implicated wrote: post #190

Earlier quoted context omitted.

As a counter to that - I've tried various flavors/quants/full weights and Qwen 3.8 27B has been entirely useless at anything non-trivial. Sure - it can do some boilerplate work (though, even armed with a well written spec and working within a very well known framework it went off the rails and did things in a way that were... um... questionable at best) but I don't see it as anything more than a personal assistant st…

I've had the exact opposite experience. I've been using 3.8 for my daily driver since last week, and I've gradually been giving it more and more complex tasks as it continues to deliver high quality results. Now I am basically handing off large complex features, and 3.8 is doing the planning, task breakdown, implementation and review with just a few notes from my side. The tradeoff is time (especially on RDMA4 hardwa…

I can't get 3.8 to exit thinking loops. It will just think and think and think on the most trivial topics. I wanted it to port a speed test powershell script to c#. Claude opus 5 completes it under 60 seconds. I let 3.8 churn about 6 different times for 30+ minutes and it never wrote a single line of code to disk. It wrote lots of lines in thinking.

unsloth/Qwen3.8-27B-GGUF UD-Q3_K_XL DSH (pi)

Any tips?

Re: GLM-5.3-Flash

#427

Post by pietz »

Kungfuturtle wrote: post #416
pietz wrote: post #141

With tiny models surpassing huge, 6 months old models on benchmarks, does anybody have some smart words to share on how these still "feel" different? Artificial Analysis ranks GPT 5.6 Luna similar to GPT 5.4, but that never matches my real world experience. AA seems to do a good job making a single number as representative as possible but there is still so much benchmarks don't communicate.

One term of art that's emerged for this "feel" is "big model smell", first coined by @aidan_mclau. [0] To my surprise I couldn't find any proper explainers of the term in a quick search, despite grokking it after seeing it in various contexts on Twitter, but Fable 5 offered a useful analogy: "A student who memorized worked solutions and one who understands the subject score the same on the test; you can only tell the…

Appreciate you taking the time. That fable analogy is well put. Almost obvious once you know it.

Re: GLM-5.3-Flash

#428

Post by saberience »

twobitshifter wrote: post #273
saberience wrote: post #226

Earlier quoted context omitted.

It's not possible to keep shrinking down parameters and keep "frontier" performance, it's like saying it's possible to take a 3 hour movie and compress it down to 3 megabytes, there are information theoretic limits on the amount of bits of information that can be compressed. What I'm saying is, if you're expecting a model that can be run on a 16GB or 32GB machine with the intelligence/knowledge of Mythos or Sol, it w…

The current models are not close to approaching the limit of compression for intelligence. They aren’t even focused on it like Chinese labs are. The training of Qwen’s 27B parameter model showed that by structuring model training from fundamentals to more difficult topics they were able to drastically reduce the number of parameters needed. The ‘frontier’ models rely on scale to achieve their results but that’s not t…

Yes they are approaching the limits, try asking smaller models niche questions about almost anything, they hallucinate massively because you cannot simply pack in all the raw knowledge from a massive frontier model into something that’s quantified down to 20GB etc.

It breaks fundamental laws of information theory. It’s like saying you can extract 100 joules of energy from 10 joules of energy source. Not possible.

Re: GLM-5.3-Flash

#429

Post by saberience »

computerex wrote: post #373
saberience wrote: post #226

Earlier quoted context omitted.

It's not possible to keep shrinking down parameters and keep "frontier" performance, it's like saying it's possible to take a 3 hour movie and compress it down to 3 megabytes, there are information theoretic limits on the amount of bits of information that can be compressed. What I'm saying is, if you're expecting a model that can be run on a 16GB or 32GB machine with the intelligence/knowledge of Mythos or Sol, it w…

You heard of JEPA? LLM's have all sorts of garbage they have memorized. Reasoning in latent space instead of in text significantly reduces the number of needed parameters.

JEPA is a joke, let me know when those models do anything useful.

Re: GLM-5.3-Flash

#430

Post by whatshisface »

KronisLV wrote: post #284
Implicated wrote: post #202

Earlier quoted context omitted.

Use Opus 4.8. 5 is absolute garbage. Don't use DS4 Flash in max effort mode. It's just spinning its wheels, in my experience (I have a harness for testing models with 25 real bugs/features/etc from my real projects that I measure outcomes against) DS4 flash does _worse_ with max effort. It will literally have the right approach and reason itself away from it.

I don’t know how 5 can do so much better in benchmarks but absolutely suck to use in practice compared to 4.X. Fable feels better, Kimi and GLM also feel better sometimes but tbh all of them make plenty of annoying mistakes.

The prompts in most of the benchmarks match what you have to write to obtain good performance from Opus 5. Reading benchmarks is very revealing.
Post reply on HN