China's latest AI model claims to be even cheaper to use than DeepSeek
Startup Z.ai, formerly known as Zhipu, announced Monday that its new GLM-4.5 AI model would cost less than DeepSeek to use. In contrast to the logic underlying existing AI models, Z.ai said its new GLM-4.5 is built on what's known as "agentic" AI, meaning that the model automatically breaks down a task into sub-tasks in order to complete it more accurately.
The new model is also open sourced, meaning it is free for developers to download and use.
At about half the size of DeepSeek's model, GLM-4.5 only needs eight Nvidia H20 chips to operate, Z.ai CEO Zhang Peng told CNBC on Monday.
That's the chip Nvidia customized for China in order to comply with U.S. export controls. The chipmaker said this month that the U.S. will allow it to resume those China sales after a three-month pause, but it's unclear when those shipments will begin.
Zhang said the company doesn't need to buy more of the chips as it has enough computing power for now, but declined to share how much Z.ai spent on training the AI model. Details will be released later, he said.
Back in January, DeepSeek had rattled global investors with its apparent ability to defy U.S. chip restrictions and create an AI model that not only rivaled U.S.-based OpenAI's ChatGPT, but also undercut it in training and operating costs.
DeepSeek claimed training costs for its V3 model were less than $6 million, although some analysts said that figure was based on the company's hardware spend of more than $500 million over time.
Z.ai said that for its new GLM-4.5 model, it would charge 11 cents per million input tokens versus 14 for DeepSeek R1; and 28 cents per million output tokens versus $2.19 for DeepSeek. Tokens are a way of measuring data for AI model processing.
Earlier this month, Alibaba-backed Moonshot released Kimi K2, which claimed to beat OpenAI's ChatGPT and Anthropic's Claude on certain coding capabilities. Kimi K2 charges 15 cents for every 1 million input tokens, and $2.50 per 1 million output tokens, according to its website.
In late June, OpenAI named Zhipu in a warning about Chinese AI progress. The U.S. has also added the startup to its entity list that restricts American companies from doing business with it.
Z.ai launched in 2019 and is reportedly planning an initial public offering in Greater China.
The startup has raised more than $1.5 billion from investors including Alibaba, Tencent and Qiming Venture Partners, according to PitchBook. Aramco-backed Prosperity7 Ventures as well as municipal funds from the cities of Hangzhou and Chengdu are also among Z.ai's backers, the database showed.
In the last few weeks, several other Chinese companies have announced new, open-source AI models. During the World AI Conference in Shanghai, Tencent released the HunyuanWorld-1.0 model for generating three-dimensional scenes for game development. Last week, Alibaba announced its Qwen3-Coder model for writing computer code.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
2 minutes ago
- Yahoo
5 Reasons Trump's Trade Deal With China Is Bad News for the Middle Class
President Donald Trump's latest trade deal with China may look like a diplomatic win, but for the American middle class, it comes with hidden costs. Trending Now: Find Out: While tariffs are being reduced in exchange for promises from Beijing, households could still face higher prices, disrupted supply chains and reduced job growth. Here are four reasons Trump's trade deal with China is bad news for the middle class and what families can do to protect their finances. Higher Consumer Prices Despite Tariff Relief Even as the U.S. and China approach an August trade deal deadline, prices on many consumer goods remain elevated, and middle-class households continue to feel the strain. Some experts argue that the new tariffs may not drastically shift average import prices. However, middle-class families are more likely to feel the impact in specific categories, such as electronics, tools and household goods. 'U.S. companies scrambled to import as many goods as possible to stockpile before new tariffs were fully implemented, mitigating the immediate impact of tariffs on prices,' said Bryan Riley, Director of the Free Trade Initiative at the National Taxpayers Union. Riley said that since imports from China account for just 13.2% of total U.S. imports, increases in the price of specific Chinese goods may not push up the overall import average. However, they can still significantly affect middle-class budgets for everyday items. Read More: Erosion of Real Incomes and Job Losses An analysis by the Federal Reserve Bank of San Francisco warned that Trump's trade measures could cut national real income by around 0.4%, while losses in services and agriculture might offset job gains in manufacturing. 'What's pitched as economic growth is actually a slow bleed: Manufacturing jobs won't magically return, and small businesses relying on predictable import costs are about to face more whiplash,' said Patrice Williams Lindo, CEO of Career Nomad. 'Wages stay stagnant while everyday costs climb. And here's the kicker — there's no workforce investment baked into this deal. That means your job security, benefits and opportunities to grow could evaporate, especially if your industry leans heavily on exports or global sourcing.' Volatile Markets and Supply Chain Instability Although the China deal eased recession fears, experts said that uncertainty around ongoing tariffs still disrupts manufacturing and logistics. Businesses may hold back investment or retool supply chains, raising costs for middle-class consumers and slowing hiring. For example, uncertainty remains one of the most significant threats to economic momentum, particularly for businesses making long-term decisions. 'The real issue is that this deal doesn't create clarity. It reinforces an environment of 'wait and see,' Robert Khachatryan, CEO and founder of Freight Right. 'That's not how you build confidence in the economy.' Khachatryan added, 'You can't expect small and midsize businesses, who employ a huge portion of America's middle class, to plan for the future when they're stuck playing defense against the next round of tariffs.' Missed Middle-Class Priorities in the Deal While the latest Trump-China deal touts manufacturing wins, some economists warn it overlooks the broader economic trade-offs that directly affect the middle class. 'We have an experiment,' said Michael Froman, president of the Council on Foreign Relations, in a recent interview on Conversations with Jim Zirin. 'In 2018, President Trump imposed 25% tariffs on steel. Seven years later, we have 1,000 more steelworkers, but 75,000 fewer workers in manufacturing sectors that relied on steel, and a 30% drop in steel sector productivity.' This kind of trade-off may deliver political wins, but it overlooks how tariff-driven policies ripple into everyday life for the middle class. 'Over time, reduced job stability in trade-sensitive sectors and a slowdown in wage growth may exacerbate economic insecurity for families already stretched thin by inflation and debt servicing costs,' said Jean-Baptiste Wautier, a private equity CIO and World Economic Forum speaker. How To Protect Your Budget Middle-class families can shield themselves by using rewards or rebate programs and strategically stockpiling essentials before potential tariff increases. Julian Merrick, founder and CEO of Supertrader, a fintech firm focused on global markets, recommends starting with a small emergency fund, even setting aside $200 to $300, which can help families avoid debt when unexpected expenses arise. 'It also helps to cut back on spending in categories where prices are rising — things like tech, clothes or imported goods,' Merrick said. 'Families should avoid taking on new high-interest debt right now, especially for non-essentials. And for those with investments, make sure the money is spread out across different industries.' Editor's note on political coverage: GOBankingRates is nonpartisan and strives to cover all aspects of the economy objectively and present balanced reports on politically focused finance stories. You can find more coverage of this topic on More From GOBankingRates 6 Hybrid Vehicles To Stay Away From in Retirement This article originally appeared on 5 Reasons Trump's Trade Deal With China Is Bad News for the Middle Class Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Bloomberg
4 minutes ago
- Bloomberg
US, China Negotiators Conclude Stockholm Trade Talks Round
US and Chinese negotiators have concluded their latest round of trade talks, conducted in Stockholm. The talks followed previous rounds in Geneva in May and in London in June.
WIRED
4 minutes ago
- WIRED
Meta's AI Recruiting Campaign Finds a New Target
Jul 29, 2025 12:10 PM Meta approached more than a dozen staffers at Mira Murati's AI startup to discuss joining its new superintelligence lab. One received an offer for more than $1 billion. Photo-Illustration:Mark Zuckerberg is on a warpath to recruit top talent in the AI field for his newly formed Meta Superintelligence Labs. After trying to gut OpenAI (and successfully poaching several top researchers), he appears to have set his sights on his next target. More than a dozen people at Mira Murati's 50-person startup, Thinking Machines Lab, have been approached or received offers from the tech giant. (Murati, for those who don't remember, was previously the chief technology officer at OpenAI.) One of those offers was more than $1 billion over a multi-year span, a source with knowledge of the negotiations tells WIRED. The rest were between $200 million and $500 million over a four-year span, multiple sources confirm. In the first year alone, some staffers were guaranteed to make between $50 million and $100 million, sources say (a spokesperson for the lab declined to comment). So far at Thinking Machines Lab, not a single person has taken the offer. Zuckerberg's initial outreach is low-key, according to messages viewed by WIRED. In some cases, he sent recruits a direct message on WhatsApp asking to talk. From there, the interviews move fast—a long call with the CEO himself, followed by conversations with chief technology officer Andrew 'Boz' Bosworth and other Meta executives. Here's a pre–Meta Superintelligence Labs recruiting message Zuckerberg sent to a potential recruit (the tone hasn't changed much today): "We've been following your work on advancing technology and the benefits of AI for everyone over the years. We're making some important investments across research, products and our infrastructure in order to build the most valuable AI products and services for people. We're optimistic that everyone who uses our services will have a world-class AI assistant to help get things done, every creator will have an AI their community can engage with, every business will have an AI their customers can interact with to buy things and get support, and every developer will have a state-of-the-art open source model to build with. We want to bring the best people to Meta, and we would love to share more about what we are building." During these conversations, Boz has been upfront about his vision for how Meta will compete with OpenAI. While the tech giant has lagged behind its smaller competitor in building cutting-edge models, it is willing to use its open source strategy to undercut OpenAI, sources say. The idea is that Meta can commoditize the technology by releasing open source models that directly compete with the ChatGPT maker. 'The pressure has always been there since the start of this year, and I think we saw that culminate with Llama 4 being rushed out of the door,' a source at Meta tells me. The rollout of Meta's latest family of models was delayed due to struggles improving its performance, and once it was released, there was a lot of drama about the company appearing to game a benchmark to make other models appear better than they actually were. Meta did not respond on the record to WIRED's request for comment prior to publication. So why weren't the flashy tactics deployed by Meta successful in recruiting TML's A-listers? Ever since Zuckerberg tapped Scale AI cofounder Alexandr Wang to colead the new lab (along with former GitHub CEO Nat Friedman), sources have been pinging me with gossip about Wang's leadership style and concerns about his relative lack of experience. Not everyone is keen to work for Wang, I'm told—though it hasn't yet stopped Zuckerberg from recruiting nearly two dozen for the lab so far (Altman said that Meta didn't get its 'top people' and 'had to go quite far down their list' at OpenAI). Other sources I spoke with say they weren't inspired by the product roadmap at Meta—money can be made anywhere, but creating what some sources see as AI slop for Reels and Facebook isn't particularly compelling. At OpenAI and Anthropic, you'll still get gobs of money, and the missions center around more lofty goals like building artificial general intelligence that 'benefits all of humanity.' Sources I spoke with who went through the interview process with Meta said it's become a way to test one's market value in the AI industry. It's not like TML is short on cash, either. The startup's latest funding round—the largest seed round in history!—suggests researchers who stay don't have to choose between being a missionary or a mercenary. The one-year-old startup is already valued at $12 billion, and it hasn't even released a product yet. Por qué no los dos , as they say. Reporting this column, I spoke to sources across most of the major AI labs to ask: Are you bullish or bearish on MSL? Rarely did I get a diplomatic 'it's too early to tell.' Instead, I heard a lot of chatter about big egos and a perceived lack of coherent strategy. For my part, and I'm not just trying to be diplomatic, I actually do think it's too early to tell. I mean, they say you can't buy taste, but that's sort of Zuckerberg's whole schtick. Now that the team officially costs Meta billions of dollars, the pressure is on to turn this recruiting sprint into a successful lab. So far, I'm told, everyone at MSL reports to Wang, and a proper org chart hasn't been finalized. Once that structure is defined, we'll be able to figure out a bit more about what the lab hopes to become—as long as Meta can hang onto its buzzy new hires. This is an edition of Kylie Robison's Model Behavior newsletter . Read previous newsletters here.
