[flagged]
Personally I'm hoping for EU to step up a bit
Seems incredibly unlikely, doesn't it? (I hope for the same thing but it feels almost like wishful thinking!)
DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
141–150 of 343 posts
Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
#142Earlier quoted context omitted.
Can't wait for the DwarfStar quants - I have been using DeepSeek v4 flash (preview) as my main coding agent for months now (running on my 128gb mbp) - it seems this model outperforms GLM 5.2 on nearly every metric. Thanks for sharing the news, I was refreshing huggingface but gave up thinking it likely would take some more time.
are you working in earplugs? :)) even with 128 gigs of ram it must be super noisy.
I actually run it as a server - so most of the time I don't have to listen to it right next to me - it's just sitting in another room in my house - but I often am traveling with it and will have it sitting right next to my coding laptop and yea the fan runs non-stop - it's not obnoxious so i can pretty easily tune it out - also airpods/noise canceling headphones help!
Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
#143Earlier quoted context omitted.
This is a straightforward false equivalency. “Western” models do not censor in the same way, nor for the same reasons, that the Chinese models do. “Just as much” is not remotely plausible, yet it’s doing all the heavy lifting.
The Anthropic and OpenAI models are much more censored and in ways that directly prevent them to be useful, e.g. by refusing to reply to elementary questions of biology and chemistry. Any normal user is much more likely to ask questions to which the Anthropic and OpenAI models do not answer, than to ask questions about the modern Chinese history, to which a Chinese LLM will not answer.
> questions about the modern Chinese history, to which a Chinese LLM will not answer
This has been debunked here on HN so many times. The Chinese open models do answer the hairy Chinese political questions, and the raw APIs pass-through the response. Now, the answer might be blocked by the agent who's calling the API, specially if you are using a Chinese endpoint instead of the RoW (i.e. Singapore) endpoint.
That's the reason why you should always prefer a open agent/harness as well instead of using the provider's.
Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
#144Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
#145The weights were just released a few minutes ago: https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731
Hmm, it's targeting a HW accelerator with a 128x128 matmul primitive, which one is that? Warp Group Matrix Multiply Accumulate on H100?
Just because scales are grouped by 128x128 tiles, does not mean you need a single compute tile that large. It works completely fine to process it with multiple smaller tiles that get given the same scales, like how this works on Hopper and Blackwell today
Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
#146[flagged]
Good luck when the last non-Chinese frontier labs will have closed and the CCP will ask to stop sharing models open source.
We'll never have fewer open-weight models than exist now. They won't suddenly disappear when labs stop publishing new ones. In fact, people will keep improving them and will keep distilling new frontier models into existing open-weight models.
Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
#147Daily reminder that none of these numbers are valid in a world where no one publishes the sampling settings used. Daily reminder that improving your samplers from the garbage default top_p/top_k to min_p or subsequent methods dramatically improves the performance of these models, and makes most quantities like measured "verbosity" and subsequent calculations of "intelligence per token" meaningless Daily reminder that…
Can you explain how sampler choice makes intelligence per token meaningless?
Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
#148Earlier quoted context omitted.
One side of a conflict being a minority does not make it less nuanced. The majority of the world is religious - doesn’t mean the debate on religion isn’t a complex question. The majority of the world approved of slavery historically. The majority of countries have ethnically cleansed their Jews, many of them in living memory. When interrogated you will find that the only ones asserting the war in Gaza is a genocide a…
Are you a part of the $1B Israel is spending to try and propagandize and rehabilitate their reputation? [1] I'm always suspicious that's the case given how mentions of gaza seem to bring out brand new accounts who only talk about Israel. [1] https://quincyinst.org/research/the-eighth-front-inside-isra...
Classic “foreign agent” ad hominem.
We are on a thread discussing Chinese models. Every discussion on here that’s negative about China or its models suddenly gets derailed via whataboutism to Israel/Gaza. A very convenient distraction.
Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
#149This seems to me like this is probably at least a large part of what OpenAI was up to yesterday with their aggressive price cutting; trying to get out in front of this.
If the full non-flash model follows up with the expected improvements, and at the price point they've been keeping, it puts the frontier labs in a tough position and it feels to me like like OpenAI is reaching deep into their pockets to try to head that off.
TFA link is a 404 though. I'm reading through https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731 instead
Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis
#150Earlier quoted context omitted.
Are you a part of the $1B Israel is spending to try and propagandize and rehabilitate their reputation? [1] I'm always suspicious that's the case given how mentions of gaza seem to bring out brand new accounts who only talk about Israel. [1] https://quincyinst.org/research/the-eighth-front-inside-isra...
Classic “foreign agent” ad hominem. We are on a thread discussing Chinese models. Every discussion on here that’s negative about China or its models suddenly gets derailed via whataboutism to Israel/Gaza. A very convenient distraction.
My history, unlike yours, is wide open. People can see I'm not a foreign agent. I don't throw this accusation at anyone other than throwaway accounts. I've had conversations in the past with pro-Israel HNs that I'd never accuse of being a foreign agent because their history is wide open and this isn't the only topic on their mind.
And yes, we were discussing censorship of models which, as I pointed out, doesn't seem like ChatGPT is directly censoring data though it does appear to be manipulating it. Pretty on topic.
It was you, brand new account hiding your past opinions, who came in here to make this solely about Israel.
Yeah, I think you are likely a foreign agent. Prove me wrong and post from an established account.