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

DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

artificialanalysis.ai

Post reply on HN
Print view View original

321–330 of 343 posts

Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

#321

Post by donquichotte »

throwa356262 wrote: post #190
donquichotte wrote: post #174

Earlier quoted context omitted.

My conspiracy theory is that this is the new space race, and the CCP encourages this to show the world what Chinese engineers are capable of, and tank the Anthropic/OpenAI valuation bubble as a desirable side effect.

Half of the engineers at OAI and Anthropic are Asian, I don't think China does all these for signaling.

I was not trying to make that kind of argument, but rather a geopolitical one.

Retaining world-class talent and home-growing industries is no small feat. In Europe, we have been struggling with this, and I am not even entirely sure why.

Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

#322

Post by smartbit »

bwfan123 wrote: post #193
L8D wrote: post #182

Earlier quoted context omitted.

If you're not paying for the model through Open router, then which provider are you buying it from? I thought that if that provider was listed on open router it would be the same price as going directly to the provider? Are you buying from a provider that isn't listed on open router already?

> Are you buying from a provider that isn't listed on open router already I am using openrouter with zdr guardrail which routes to any provider that supposedly doesnt train on user data. I also use fireworks (directly, not via openrouter) which is a provider promising zdr and has a bunch of open weights models. My issue is that these zdr providers dont transparently disclose caching/tokens etc and so they end up bein…

> OpenRouter has taken the stance that in-memory caching of prompts is not considered “retaining” data, and we therefore allow endpoints/models with implicit caching to be hit when a ZDR routing policy is in effect. [0]

OpenRouter doesn’t list DeepSeek as ZDR combined with Caching. Thanks for the tip of Fireworks as Fireworks has Zero Data Retention by default. [1]

[0] https://openrouter.ai/docs/guides/features/zdr

[1] https://docs.fireworks.ai/guides/security_compliance/data_ha...

Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

#323

Post by dgellow »

UltraSane wrote: post #317
dgellow wrote: post #273

Earlier quoted context omitted.

Im not sure if you’re joking or not. - the US strictly regulate cryptography https://en.wikipedia.org/wiki/Export_of_cryptography_from_th... - some prime numbers are considered illegal https://en.wikipedia.org/wiki/Illegal_number#Illegal_primes I don’t think banning open weight is that far fetch compared to those

Neither of those worked very well.

Ok, but that’s a different thing altogether. I personally expect a ban, that doesn’t mean it will be very effective, but that does mean it will have a meaningful impact (such as not being able to offer open weight models as a service)

Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

#324

Post by swingboy »

willchis wrote: post #102
ComputerPerson wrote: post #82

Earlier quoted context omitted.

Is the difference between this and a frontier model that the scripture is guaranteed to be real? I'm on a team that develops a Bible study app, and we're all relatively content with how the basic models converse regarding scripture. Even as far back as GPT-4 was excellent. They occasionally have minor hallucinations (a dealbreaker for a production app), but they do an excellent job with theology and Bible scholarship…

I hope the irony of being concerned about hallucinations in a bible study app is not lost on you or others reading this!

Sir, you dropped your fedora.

Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

#325

Post by Bayart »

0cf8612b2e1e wrote: post #106

Somewhat relatedly, how do the economics for Huggingface work? They must be hosting petabytes of models and datasets by now. I have downloaded quite a few “just in case”, only to replace them with the later iteration months later. Does the file hosting actually cost peanuts when you do it yourself and the cloud has shattered my understanding of what it actually costs to deliver so much data?

Nowadays storage and bandwidth is cheap as long as you're not going through a hyperscaler and it doesn't require compute (ie. "dead" data). I'm working mostly on GCP so I'm not sure what are the rates now, but at least AWS were known for billing insane numbers on egress.

As long as you're smart on caching on the edges and deduplication/overlaying with your content (which structured data types like models, containers and repos typically are fit for) you can get remarkably far for less than you'd think.

The economics of serving models seem to be way more dodgy.

Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

#326

Post by Bayart »

flyingpenguin wrote: post #281
oceanplexian wrote: post #232

Earlier quoted context omitted.

It's less than $1/TB. If you have settlement-free peering it's like fractions of a penny at scale. Cloud provider DTO pricing is literally the biggest scam on Earth.

My company spends $8,000,000/month on storage and $4,000,000/month on data with aws... In the financial reports we do we round this number away. Make of that what you will

"Nobody ever got fired for buying AWS."

Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

#328

Post by pornel »

tmikaeld wrote: post #78

My problem with DS flash/pro is that they don’t push back on obvious bullshit, both irl and code [0] but it’s a great implementer workhorse if you give it _very_ detailed specs. [0] https://petergpt.github.io/bullshit-benchmark/viewer/index.v...

omp.sh lets you enable "advisor" model that watches the session. Works very well for keeping DS on track.

Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

#329

Post by jmathai »

ComputerPerson wrote: post #130
jmathai wrote: post #126

Earlier quoted context omitted.

Philosophically, I don’t believe we should outsource the accuracy of scripture to any single entity (let alone a for profit secular one). So it’s less about model choice and more about governance of scripture. I will check out the link you sent for sure!

I was under the impression you were getting BOOK.CHAPTER.VERSE references from the model and then sourcing them from a ground truth db. Anyways, impressive app! We haven't tackled such an ambitious project just for it being daunting.

Took a look at StepBible. Looks really cool. My email is in my profile if you wanted to connect.

Re: DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

#330

Post by xXSLAYERXx »

Danox wrote: post #310
xXSLAYERXx wrote: post #258

Earlier quoted context omitted.

You're clearly not an American and they will never catch up on hardware. I'm mostly curious why you are so excited for China to "crush" American companies?

Never?… Ha Ha Ha

Right. I misspoke. Should have said not in our lifetime.
Post reply on HN