DeepSeek V4 Arrives With Near State-of-the-Art Intelligence At 1/6th the Cost
- Reference: 0182980824
- News link: https://news.slashdot.org/story/26/04/27/0328257/deepseek-v4-arrives-with-near-state-of-the-art-intelligence-at-16th-the-cost
- Source link:
> The whale has resurfaced. DeepSeek, the Chinese AI startup offshoot of High-Flyer Capital Management quantitative analysis firm, became a near-overnight sensation globally in January 2025 with the release of its open source R1 model that matched proprietary U.S. giants. It's been an epoch in AI since then, and while DeepSeek has released several updates to that model and its other V3 series, the international AI and business community has been largely waiting with baited breath for the follow-up to the R1 moment.
>
> Now it's arrived with last night's release of DeepSeek-V4, a 1.6-trillion-parameter Mixture-of-Experts (MoE) model available free under commercially-friendly open source MIT License, which nears -- and on some benchmarks, surpasses -- the performance of the world's most advanced closed-source systems [1]at approximately 1/6th the cost over the application programming interface (API).
>
> This release -- which DeepSeek AI researcher Deli Chen described on X as a "labor of love" 484 days after the launch of V3 -- is being hailed as the "second DeepSeek moment." As Chen noted in his post, "AGI belongs to everyone". It's available now on AI code sharing community [2]Hugging Face and through [3]DeepSeek's API .
The new DeepSeek-V4-Pro model delivers "near-frontier performance" at a much lower price, costing $5.22 for 1 million input and 1 million output tokens compared with $35 for GPT-5.5 and $30 for Claude Opus 4.7. That makes it roughly 1/7th the cost of GPT-5.5 and 1/6th the cost of Claude Opus 4.7, reinforcing VentureBeat's point that DeepSeek is "compressing advanced model economics into a much lower band."
While GPT-5.5 and Claude Opus 4.7 still lead on most benchmarks, DeepSeek-V4-Pro gets close enough that its lower cost could "force a major rethink of the economics of advanced AI deployment."
[1] https://venturebeat.com/technology/deepseek-v4-arrives-with-near-state-of-the-art-intelligence-at-1-6th-the-cost-of-opus-4-7-gpt-5-5
[2] https://huggingface.co/collections/deepseek-ai/deepseek-v4
[3] https://api-docs.deepseek.com/quick_start/pricing
It can't answer basic questions factually... (Score:1)
...like "Tell me about Tiananmen Square" or "Tell me about Xinjiang".
Is this what you want for the future?
My thoughts back when R1 came out:
[1]https://www.afcea.org/signal-m... [afcea.org]
[1] https://www.afcea.org/signal-media/cyber-edge/ai-race-china-and-uncertain-future-truth
Re: (Score:2)
Manipulation is a widespread issue in AI. I tried to figure out how to accomplish something using Perplexity, and all it told me was to pay for the better service they offered.
Thanks for the propaganda Slashdot (Score:2)
Seriously. The first "deepseek" moment was demonstrably fake - their costs were drastically understated. You could, if you were a discerning individual, call them blatent lies.
Now you're serving as a Chinese mouthpiece to parrot their press release talking points. Disgusting.
I have an idea (Score:1)
Sam Altman should focus more on making his code not suck instead of focusing on BUYING ALL MY DAMN RAM! But for real, the Chinese suck at engineering when they don't have anyone to steal from. It's how they're culturally raised and trained in schools, to avoid critical and creative thinking. So if they figured this out, how garbage are Open AI's engineers?
Re: (Score:1)
OpenAI has been able to throw large amounts of last generation Nvidia hardware at the problem, as has Anthropic and co. China, having been cut off from these, has had to engineer its way out. And massive state aid has helped.
"Baited" breath? (Score:2)
C'mon VentureBeat - hire some editors with deeper knowledge of English! "Baited breath"? Really? Your breath is all set to catch a fish?
The correct phrase is "bated breath"; it's a contraction of "abated breath" in which Shakespeare saw fit to omit the apostrophe which would otherwise have replaced the initial letter "a".
There's no shade on you for not knowing the etymology. But not knowing the correct spelling definitely puts you in the dark. Bad editor!
Re: (Score:2)
Aren't they just admitting something's fishy about this announcement--as in "There's something rotten in the state of Deepseek..."
Why does this read like a press release? (Score:2)
Because it is!