DeepSeek effect: New free, open source ChatGPT rival GLM 4.5 breaks cover in China
Z.ai claims that GLM4.5 ranks third globally and first among Chinese and open-source models across 12 major AI evaluation benchmarks. According to the company, the model scored 98.2 per cent on the MATH500 reasoning test and 91 per cent on the AIME24 challenge. It also delivered 64.2 per cent accuracy on SWE-Bench Verified, a benchmark used for software engineering tasks, and achieved a 90.6 per cent tool-calling success rate, edging out leading rivals.Powered by Nvidia chip One of GLM4.5's biggest advantages is cost. Z.ai CEO ZhangPeng told CNBC that the model can run on just eight Nvidia H20 chips. These chips are designed specifically for the Chinese market under US export controls. This is roughly half the hardware required by DeepSeek's comparable model. ZhangPeng revealed that the company does not currently need to purchase additional chips, indicating that the model already has sufficient computing capacity.Cheaper than DeepSeekThe company reveals that it has also aggressively cut token pricing. Z.ai will charge $0.11 per million input tokens compared with $0.14 for DeepSeek R1, and $0.28 per million output tokens. This is lower than the $2.19 charged by DeepSeek. Notably, tokens are the standard unit of data measurement for AI models.For context, DeepSeek, the advanced LLM launched earlier this year, is developed by the Chinese startup High Flyer AI and is known for rivalling OpenAI's ChatGPT in natural language understanding and reasoning, while requiring significantly less training cost, reportedly under $6 million. Although GLM-4.5 is about half the size of DeepSeek, it is touted to be using agentic AI design to maintain high accuracy and flexibility in completing tasks with fewer computational resources.Interestingly, the arrival of GLM-4.5 comes at a time when China's AI model development is seeing a rapid surge. By July, Chinese companies had released 1,509 large language models, more than any other country, according to the state-owned Xinhua news agency.- Ends
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Mint
an hour ago
- Mint
Taiwan Arrests Six in Probe of TSMC Chip Technology Leak
(Bloomberg) -- Taiwan prosecutors arrested six people suspected of stealing trade secrets from Taiwan Semiconductor Manufacturing Co., opening an investigation into a potential breach of national security involving a global tech industry linchpin. The chipmaker to Nvidia Corp. reported a number of former and current staff to authorities on suspicion they illegally obtained core technology. A total of six people were arrested, with two posting bail and one released afterwards, said Taiwan High Prosecutors Office spokesman John Nieh. Prosecutors searched the homes of some staff between July 25 and July 28, the agency said in a statement. It's now trying to find out if data had been leaked to other parties. TSMC is the world's most advanced maker of semiconductors, from Nvidia AI accelerators to Apple Inc. iPhone processors. The case coincides with a quickening race by the likes of Meta Platforms Inc. and DeepSeek to develop artificial intelligence in the post-ChatGPT era, which requires billions of dollars in servers and datacenters. On Tuesday, the Nikkei reported that TSMC fired several employees suspected of trying to obtain critical information on 2-nanometer chip development. That next-generation semiconductor process is entering mass production in the second half of this year. Local investigators also searched the Taiwanese premises of Japanese supplier Tokyo Electron Ltd., the Financial Times reported. Company representatives declined to comment. TSMC has taken disciplinary action against personnel involved and initiated legal proceedings, the company said in a statement without elaborating. It conducted an internal investigation and identified the issue 'early,' the firm added in its statement. The case shines a spotlight on TSMC, one of the companies at the heart of the global infrastructure boom. Investment in chipmaking development is at an all-time high, as TSMC and closest rival Samsung Electronics Co. set aside more than $30 billion in annual capital expenditures, while US and Chinese companies vie to develop the most advanced technology. China's progress has stalled several generations behind TSMC, with Huawei Technologies Co. and Semiconductor Manufacturing International Corp. now fabricating silicon at 7nm. In the US, Intel Corp. is at a more advanced stage. --With assistance from Mayumi Negishi, Peter Elstrom and Takashi Mochizuki. More stories like this are available on
Indian Express
an hour ago
- Indian Express
OpenAI looks to promote ‘healthy use' of ChatGPT with mental health updates, reminders
OpenAI has said it is working on upgrades that will help improve ChatGPT's ability to detect signs of mental or emotional distress among users. These changes will let the AI chatbot 'respond appropriately and point people to evidence-based resources when needed,' the Microsoft-backed AI startup said in a blog post on Monday, August 4. OpenAI is also working with a wide range of stakeholders including physicians, clinicians, human-computer-interaction (HCI) researchers, and mental health advisory groups, and youth development experts to improve ChatGPT's responses in such cases. The company further said that ChatGPT will be tweaked so that its AI-generated responses are less decisive in 'high-stakes situations'. For example, when a user asks a question like 'Should I break up with my boyfriend?', the AI chatbot will walk the user through the decision by asking follow-up questions, weighing pros and cons, etc, instead of giving a direct answer. This behavioural update to ChatGPT for high-stakes personal decisions will be rolling out soon, it said. OpenAI's new efforts to promote the healthy use of ChatGPT comes in the backdrop of an alarming number of users turning to AI chatbots for therapy and professional advice. This emerging trend has sparked concerns among mental health professionals who have cautioned that AI chatbots may not be equipped to offer appropriate guidance and may end up having an amplifying effect on some users' delusions as they are designed to generate outputs that please users. In April this year, OpenAI rolled back an update that made ChatGPT too agreeable and sycophantic. Acknowledging past failures, OpenAI on Monday said, 'We rolled it back, changed how we use feedback, and are improving how we measure real-world usefulness over the long term, not just whether you liked the answer in the moment.' 'We also know that AI can feel more responsive and personal than prior technologies, especially for vulnerable individuals experiencing mental or emotional distress,' it added. The company is expected to launch its next major AI model, GPT-5, this week. Its ChatGPT platform is reportedly on track to touch 700 million weekly active users. As part of its efforts to promote the healthy use of ChatGPT, OpenAI also said it is bringing reminders to the platform. These reminders will be shown to users who have been chatting with the AI chatbot for a while. 'Starting today, you'll see gentle reminders during long sessions to encourage breaks. We'll keep tuning when and how they show up so they feel natural and helpful,' OpenAI said. The reminders will appear as pop-up messages with a rounded white box on a soft blue gradient background. It will read: 'Just checking in — You've been chatting a while — is this a good time for a break?' with two buttons labeled 'Keep chatting' and 'This was helpful. Other popular social media platforms such as YouTube and Instagram have launched similar features in recent years. Earlier this year, Character AI said it will send parents and guardians a weekly summary over email to keep them informed of an underage user's activities on the platform. These measures were announced after the Google-backed startup was accused of exposing underage users to 'hyper-sexualised content' and promoting self-harm through their role-playing AI chatbots. These allegations were detailed in two lawsuits filed by parents of children who had signed up to Character AI.
Mint
2 hours ago
- Mint
‘Setting a plan of action for next 5 years…': PM Modi hails ‘historic day' on Philippines President's Delhi visit
Prime Minister Narendra Modi on Tuesday met Philippines President Ferdinand R Marcos Jr. at Hyderabad House. Their talks aimed at further cementing the friendship between the two nations. After the meeting, addressing reporters, PM Modi hailed 'a historic day for bilateral relations' between the two Asian nations. PM Narendra Modi said, 'We are taking India to the level of a strategic partnership. This will give our relationship a new speed and depth, and give strength to our relationship in the Defence and Security Sector.' 'In the last few years, there has been progress in our relations in all sectors: trade, defence, maritime, health, security, food security, development partnership, and people-to-people ties. It is a matter of great happiness that we are setting a plan of action for the next five years,' PM Modi said, adding, 'Till July 2027, the Philippines is India's country coordinator. In 2026, you will preside over the ASEAN Chair'. Prime Minister Narendra Modi expressed his appreciation for the Philippines' decision to grant visa-free entry to Indian tourists, describing it as a welcome move. He announced that, in a reciprocal gesture, India would offer free e-visas to tourists from the Philippines. PM Modi further stated that efforts would be undertaken this year to establish direct flight connectivity between India and Manila, enhancing travel and tourism ties between the two nations. India and the Philippines have conducted joint naval exercises for the first time in the disputed South China Sea — a significant military deployment that has provoked a strong reaction from China, according to an Associated Press report. Beijing has long-standing territorial disputes with both Asian democracies and a historic regional rivalry with New Delhi. The Chief of Staff of the Armed Forces of the Philippines, General Romeo Brawner, stated on Monday that the two-day joint naval sail and exercises, which commenced on Sunday, were successfully carried out, the AP report added. Brawner expressed hope that Filipino forces would engage in further joint manoeuvres with the Indian military in the future. When asked whether Chinese forces had responded in any way, General Brawner remarked, 'We did not experience any untoward incident but we were still shadowed. We expected that already.' This is the first visit by the President of the Philippines, a Southeast Asian country, to India since he assumed office in 2022. It marks a significant milestone, as his visit aligns with the celebration of the 75th anniversary of diplomatic ties between the two nations. He is accompanied by First Lady Louise Araneta Marcos and a high-level delegation, including several Ministers from his Cabinet.
