Live data from Hacker News

DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis

artificialanalysis.ai

311–320 of 343 posts

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

#311
post #98
post #37

here it is on openrouter https://openrouter.ai/deepseek/deepseek-v4-flash-0731

Half OT: Why do the cache hit rates seem to vary so much between harnesses? I use pi, which is very minimalist, and I get a hit rate of ~99%. Paying like $1 a day for Flash. Yet, the hit rate mentioned on OpenRouter is only ~79%.

Some harness (cc) injects system messages in the middle, this kills the cache as the model provider moves the system to the top and concatenate them together. Use homemade proxy, you will get same 99% hot rate with even cc.

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

#312
I got a disappointing pelican on default reasoning mode: https://static.simonwillison.net/static/2026/deepseek-flash-...

But a REALLY good pelican on reasoning mode high (via OpenRouter): https://static.simonwillison.net/static/2026/deepseek-flash-...

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

#313
What's most impressive here is that that DeepSeek keeps showing how much of a performance improvement can come from post-training alone while the architecture stayed the same. It's a strong reminder that we're probably underestimating how much optimization is still left after pretraining

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

#314
post #303
post #20

Earlier quoted context omitted.

I’ve been using v4 flash for an app I’m building [1] and it’s amazing how cost effective and good it is coming from having always used gpt, opus and sonnet models. It’s so cost effective I can offer a generous free tier since my goal isn’t to make money with it. [1] https://trysojourn.app

have you used it for anything outside of coding?

I tried using Deepseek with Claude Code and it under performed. The use I was referring to was within a product and I feel it's a lot of value for the cost.

https://api-docs.deepseek.com/quick_start/agent_integrations...

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

#315

Earlier quoted context omitted.

Chinese are pretty pragmatic. Even if they produce more expensive chips and memory, they are still going to focus on value. But who knows, maybe India will step up again like they did for Y2K 3 decades ago.

[flagged]

I said absolutely nothing about Americans in my comment so I’m a bit confused by your comment.

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

#316

Earlier quoted context omitted.

But why should I care? If my metrics are speed and cost? How many tokens it takes as a user is arbitrary to some extent.

If speed is a metric for you, tokens required to solve a problem affects that metric. All else being equal passing triple the amount of tokens through a model to solve the same problem makes it slower. Doesn’t mean this model is bad, and it has to be considered how cheap it is, but it’s a factor.

Still, seems useless to measure tokens when you can measure speed.

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

#317

Earlier quoted context omitted.

Don't be glib, the 1st amendment is taken VERY seriously.

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.

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

#318
post #37

here it is on openrouter https://openrouter.ai/deepseek/deepseek-v4-flash-0731

Am I parsing the providers correctly that if I want a US-based zdr provider that offers the best price and performance that is currently Cloudflare and no longer Fireworks (which is that it was with the previous version)?

What am I missing?

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

#319

Earlier quoted context omitted.

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…

Not the person you're replying to. I'm white/non Jewish, and I have a law degree and work in the legal industry. My post history is 'wide open'.

Under no definition of genocide (as defined either going back to the middle 20th century or as that term is defined legally today) is Israel committing genocide. Simply put, there is no evidence that Israel is committing "specific acts with the intent to destroy, in whole or in part, a national, ethnical, racial, or religious group." Intent is, as with almost everything in the criminal law, everything here. If you want to subdivide Muslimes in a way that results in, say, Hamas and Hezbollah being their own religious groups then you could mount the argument on that basis, but those are political designations and so good luck with that. And of course neither of those groups encompass the entirety of the national polities they claim to represent - just look at how hated Hezbollah is in Lebanon. They are being ostracised and even the Lebanese leaders want nothing to do with them at this point.

People like you degrade actual acts of genocide such as what happened with the Nazis or such as what has happened several times in SE Asia and Africa over the past 50 years. A country acting in self-defense is not genocide no matter how much it might hurt your sensibilities. Genocide is, as is obvious to most people, extremely difficult to prove. See Armenia for how fraught this debate is. But acts of brutality in a conflict or an extreme imbalance of military power between belligerents alone do not mean that genocide is occurring.

Good day.

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

#320
post #20

If deepseek v4 flash is beating DeepSeek V4 Pro, can we expect new V4 Pro which is on par with Opus 5 in couple weeks (even better if it beats Opus)?

I’ve been using v4 flash for an app I’m building [1] and it’s amazing how cost effective and good it is coming from having always used gpt, opus and sonnet models. It’s so cost effective I can offer a generous free tier since my goal isn’t to make money with it. [1] https://trysojourn.app

[dead]
Post reply on HN