UAE Enters Global AI Race with Powerful, Efficient Reasoning Model K2 Think

The United Arab Emirates Releases a Tiny But Powerful AI Model

K2 Think compares well with reasoning models from OpenAI and DeepSeek but is smaller and more efficient, say researchers based in Abu Dhabi. The United Arab Emirates (UAE) has released an open source model that performs advanced reasoning as well as the best offerings from both the United States and China -- one of the strongest signs so far that the nation's big investments in artificial intelligence are starting to pay off. The new model, K2 Think, comes from researchers at Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) located in UAE's capital Abu Dhabi. The model -- one of the first so-called "sovereign" AI models that incorporates technical advances needed for reasoning -- is being made available for free by G42, an Emirati tech conglomerate backed by Abu Dhabi's sovereign wealth funds. G42 is running the model on a cluster of Cerberas chips, an alternative to Nvidia's hardware. K2 Thinking is one of the UAE's contributions to the global race to demonstrate prowess in a technology widely expected to have huge economic and geopolitical implications. The United States and China are considered the dominant players in this contest. But many smaller nations, especially ones with considerable wealth to invest, are also racing to develop their own "sovereign" AI models. K2 Think is relatively modest in size, with 32 billion parameters. It is not a complete large language model but rather a model specialized for reasoning, capable of answering complex questions through a simulated kind of deliberation rather than quickly synthesizing information to provide an output. For such tasks, the researchers say it performs on par with reasoning models from OpenAI and DeepSeek, which have more than 200 billion parameters. "This is a technical innovation or, in my opinion, a disruption," Eric Xing, MBZUAI's president and lead AI researcher told WIRED ahead of today's announcement. Xing says the model demonstrates a particularly effective combination of a number of recent technical innovations. These include fine tuning on long strings of simulated reasoning; an agentic planning process that breaks problems down in different ways; and reinforcement learning that trains the model to reach verifiably correct answers. Other innovations allow the model to be served very efficiently on Cerebras chips. "How to make a smaller model function as well as a more powerful one -- that's a lesson to learn, if other people want to learn from us," Xing said. Xing adds that K2 Think was developed using several thousands of GPUs (he declined to give a precise number), and the final training run involved 200-300 chips. The plan is to incorporate K2 Think into a full LLM in the coming months. MBZUAI has open sourced the model and published a technical report that details how different innovations were combined to create it. Other nations in the Middle East, including Saudi Arabia, are also investing heavily in AI infrastructure and research. President Donald Trump traveled to the region in May to announce numerous AI deals involving US tech companies. The UAE's leadership has invested billions to establish itself as a strategically important research hub. The country has already revealed some cutting edge AI research and established an outpost in Silicon Valley. The UAE has lessened its ties to China in return for access to the US silicon needed to train frontier models. Peng Xiao, CEO of G42, and a MBZUAI board member, said in a statement: "By proving that smaller, more resourceful models can rival the largest systems, this achievement shows how Abu Dhabi is shaping the next wave of global innovation."

Abu Dhabi launches low-cost AI reasoning model in challenge to OpenAI, DeepSeek

"What we're discovering is that you can do a lot more with less," Richard Morton, managing director for the UAE's AI-focused university MBZUAI, told CNBC. A new challenger in the global artificial intelligence race has entered the ring. The Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), an AI-focused research university established by the United Arab Emirates, announced on Tuesday the release of a new, low-cost reasoning model to rival OpenAI and DeepSeek. It comes after DeepSeek, a Chinese AI lab, earlier this year shocked the world with the release of a reasoning model called R1 which it said could outperform OpenAI but with far less training costs. At just 32 billion parameters, MBZUAI's model, dubbed K2 Think, is much smaller than competing systems from OpenAI and DeepSeek. It was built on top of Alibaba's open-source Qwen 2.5 model and is run and tested on hardware provided by AI chipmaker Cerebas. For context, DeepSeek's R1 has a total of 671 billion parameters, which is essentially another term for the variables that an AI language model learns to understand and generate language. OpenAI doesn't disclose the parameter counts of its AI models.

United Arab Emirates Joins U.S. and China in Giving Away A.I. Technology

Cade Metz has covered artificial intelligence for more than 15 years. In a move that shows the growing influence of the United Arab Emirates in the global artificial intelligence race, a new research lab backed by the Persian Gulf nation said on Tuesday that it was freely sharing an A.I. model meant to compete with systems released by companies in the United States and China. Over the past year, many Chinese companies have aggressively shared their technologies through a process called open source, hoping to undercut leading U.S. companies like OpenAI and Google. Last month, OpenAI freely shared two of its own models in an effort to level the playing field and ensure that the world's software developers and businesses continued to use its technology. Now the new Emirati lab, the Institute of Foundation Models, has released its first open source model, K2 Think. The lab said the system performed on a par with the leading open source technologies from OpenAI and China's DeepSeek, according to standard benchmarks. The Emirates is among several nations pouring billions of dollars into computer data centers and research to compete with leading nations like the United States and China in artificial intelligence. Countries such as Saudi Arabia and Singapore are embracing the idea that the A.I. is so important, each should have its own version of the technology. "A.I. will not be monopolized by just a few countries," said Eric Xing, president of the Mohamed bin Zayed University of Artificial Intelligence, which operates the Emirates' new lab. "We are trying to build a foundation for research and development and sovereignty of intellectual creativity in this country." The new open source release from the Emirates is likely to divide opinion in Washington. The lab built its technology using data centers operated by G42, an Emirati firm that recently received more than 10,000 computer chips from the United States as part of a deal between the Trump administration and the Emirates. The Trump administration officials who drove the deal, including David Sacks, the White House A.I. czar, have championed the agreement as a way of persuading gulf states to use and promote American A.I. technology rather than turning to China. But as the release of the K2 Think model shows, the Emiratis are also competing with their American counterparts. "This is what every technology challenger does: They open source so they can level the playing field," said Pablo Chavez, an adjunct senior fellow with the Center for a New American Security, a think tank, who has written about A.I. infrastructure. The Institute of Foundation Models was founded in March by the Mohamed bin Zayed University of Artificial Intelligence, a six-year-old university. The university is part of an Emirati strategy called Artificial Intelligence 2031, which seeks to "position the country as a global A.I. leader." The university is led by Dr. Xing, a Chinese-born American researcher who was a professor of computer science at Carnegie Mellon in Pittsburgh. He and the Emirati university recently opened satellite offices in Paris and Silicon Valley. Dr. Xing said in an interview that the lab had built its new model with only about 2,000 specialized computer chips, far fewer than the hundreds of thousands available to leading companies in the United States. Like DeepSeek, which demonstrated its success early this year, the lab aims to show that powerful technologies can be built without access to the enormous amounts of computer hardware amassed by the likes of OpenAI and Google. "Going big, going expensive, may work, but it may not be the only way," Dr. Xing said. "We can use limited resources to make things work." K2 Think is designed for areas like math, computer coding and science research; it is not intended to be a general-purpose chatbot like ChatGPT. K2 Think is a "reasoning model" that can spend time "thinking" through complex problems before settling on an answer. Even as the Emirates works to build powerful A.I. with fewer resources, its resources are expanding, with U.S. chipmakers like Nvidia and the start-up Cerebras supplying the country with hardware. At the same time, U.S. companies are turning to the Emirates and other Middle Eastern countries for funding and other resources. In May, OpenAI unveiled plans to build a massive computing complex in the Emirates in a joint venture with G42, the software giant Oracle, the chipmaker Nvidia and others. (The New York Times has sued OpenAI and its partner, Microsoft, claiming copyright infringement of news content related to A.I. systems. OpenAI and Microsoft have denied those claims). G42 is also expected to contribute money to the construction of OpenAI data centers in the United States. For every dollar that the firm and its partners invest in the Emirates, they will invest an equivalent amount in the U.S. data centers, according to OpenAI. The size of these planned data centers suggests that G42 will invest tens of billions of dollars in each country, as the Emirates expands its influence in the A.I. field and diversifies its investments beyond oil and gas. Last week, the Qatar Investment Authority invested in Anthropic, another prominent A.I. start-up in the United States, as part of a funding round that brought $13 billion into the company. The Qatari wealth fund had already invested in Elon Musk's artificial intelligence start-up, xAI.

UAE Lab Releases Open-Source Model to Rival China's DeepSeek

The United Arab Emirates hopes to compete with the U.S. and China in the global AI race. The United Arab Emirates wants to compete with the U.S. and China in AI, and a new open source model may be its strongest contender yet. An Emirati AI lab called the Institute of Foundation Models released K2 Think on Tuesday, a model that researchers say rivals OpenAI’s ChatGPT and China’s DeepSeek in standard benchmark tests. “With just 32 billion parameters, it outperforms flagship reasoning models that are 20x larger,†the lab wrote in a press release on Tuesday. DeepSeek’s R1 has 671 billion parameters, though only 37 billion are active. Meta’s latest Llama 4 models range from 17 billion to 288 billion active parameters. OpenAI doesn't share parameter information. Researchers also claim that K2 Think leads “all open-source models in math performance†across several benchmarks. The model is intended to be more focused on math, coding, and scientific research than most other AI chatbots. The Emirati lab’s selling point for the model is similar to DeepSeek’s strategy that disrupted the AI market earlier this year: optimized efficiency that will have better or the same computing power at a lower cost. "By proving that smaller, more resourceful models can rival the largest reasoning systems, this milestone marks the beginning of the next wave of AI innovation," Peng Xiao, council member of Abu Dhabi's AI and Advanced Technology Council, said in the press release. Xiao is also the CEO of G42, an Emirati AI development company that co-launched the K2 Think model. The company last made headlines for inking a multibillion-dollar data center deal with the Trump administration earlier this year, which has since been mired in national security concerns. The lab is also aiming to be transparent in everything, “open-sourcing not just models but entire development processes†that provide “researchers with complete materials including training code, datasets, and model checkpoints,†IFM said in a press release from May. The long-term plan is to incorporate K2 Think into a full LLM in the coming months, WIRED reported. The Emiratis are serious about AI. The country counted 672 new AI companies between June 2023 and June 2024, making it the fastest-growing AI cluster in the Middle East and North Africa region. The lab behind K2 Think was established by the Mohamed bin Zayed University of Artificial Intelligence, which has its headquarters in Abu Dhabi and two research hubs in Silicon Valley and Paris. It was founded a couple of years ago as part of the Emirati government’s AI overhaul strategy, called Artificial Intelligence 2031. The university’s president is Chinese-born American researcher Dr. Eric Xing. The Emirati AI overhaul has also benefited American companies. Abu Dhabi state AI fund MGX, chaired by the UAE's national security adviser, is a founding partner of Trump's Project Stargate, and the fund has previously invested in OpenAI. Trump also announced in May that Abu Dhabi and Washington were partnering to create the largest AI data center cluster outside of the U.S., built and operated by G42 â€"the company that co-launched K2 Think on Tuesdayâ€" and with the help of Nvidia, OpenAI, Cisco, Oracle, and Japanese firm Softbank. But that deal has faced U.S. regulator scrutiny over security concerns, particularly regarding the UAE’s relationship with China. Chinese tech and AI firms like Huawei and Alibaba have been expanding their influence in the UAE and the Middle East at large. G42 specifically had several Chinese investments, which the company reportedly got rid of after pressure from the Biden administration over a $1.5 billion strategic investment by Microsoft. The Emirates’ growing bet on AI is driven by a desire to diversify investments and reduce its economic dependence on fossil fuels like oil and gas. The trend is seen in the rest of the Arab world as well, particularly in the oil-rich Gulf states. Saudi Arabia created a $100 billion AI investment fund last year as part of an effort to diversify its oil-reliant economy by 2030. Saudi Arabian DataVolt is moving forward with plans to invest $20 billion in AI data centers and energy infrastructure in the United States. Elsewhere in the Gulf, Qatar Investment Authority was a significant investor in American AI company Anthropic in a $13 billion funding round last week. A growing list of countries is seeking to join the battle for global AI dominance that's currently dominated by American and Chinese companies. In the rest of Asia, Singapore is a rising power, with AI-friendly regulatory oversight that has spurred the launch of AI innovation hubs from tech giants like Microsoft. In Europe, the French are making major AI plays. The French AI startup Mistral just secured a $1.5 billion investment by Dutch semiconductor maker ASML on Tuesday. Mistral is considered a major competitor to OpenAI in Europe.

UAE launches new low-cost AI model, challenging OpenAI and DeepSeek

A new, cheaper artificial intelligence (AI) model has entered the technology race, this time from the United Arab Emirates (UAE). The Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) in Abu Dhabi on Tuesday unveiled the release of a low-cost reasoning model that it hopes will rival DeepSeek and OpenAI. In January, the AI bubble got a shot of air after China-based research lab DeepSeek said it had been closely catching up to the achievements of the United States's OpenAI, which makes ChatGPT, using a fraction of its budget and energy. The UAE's model, called K2 Think, is smaller in terms of parameters, or the configuration variables of a machine learning model which control how it processes data and makes predictions, compared to its AI competitors, including DeepSeek. However, the researchers behind it say its performance is on par with OpenAI and DeepSeek's reasoning models. The university said in a press release that its K2 Think is "a new class of reasoning model," adding that "it employs long chain-of-thought supervised fine-tuning to strengthen logical depth, followed by reinforcement learning with verifiable rewards to sharpen accuracy on hard problems". "What was special about our model is we treat it more like a system than just a model," Hector Liu, director of MBZUAI's institute of foundation models, told CNBC in an interview. "So, unlike a regular open source model where we can just release the model, we actually deploy the model and see how we can improve the model over time". MBZUAI also said that it is "one of the fastest and most efficient reasoning systems in existence". It says K2 Think can achieve 2,000 tokens per second, which is roughly 1,500 words. K2 Think was built on Alibaba's Qwen 2.5 large language model and is run on hardware from AI chipmaker Cerebras. Like DeepSeek's R1 model, K2 Think is also open source, meaning its training data and weights are available to the public. "This new level of transparency ensures that every step of how the model learns to reason can be studied, reproduced, and extended by the global research community," the university said. The technology could have major implications for the global AI race. Though the US, followed by China, reigns supreme, other countries are trying to make their mark in AI. "K2 Think is a defining moment for AI in the UAE," MBZUAI said. "It reflects how open innovation and close public-private partnerships can position Abu Dhabi as a global leader in AI, demonstrating that the future of reasoning will be shaped not only by size, but by ingenuity and collaboration".

MBZUAI unveils K2 Think reasoning model based on Qwen 2.5

The Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) in Abu Dhabi has unveiled K2 Think, a low-cost reasoning model designed to rival systems developed by DeepSeek and OpenAI. The launch underscores the UAE's ambitions to become a significant player in the global AI sector, currently dominated by the United States and China. MBZUAI describes K2 Think as "a new class of reasoning model." Instead of relying on massive parameter counts, the system emphasizes efficient architecture and logical depth. According to the university, it uses long chain-of-thought supervised fine-tuning to improve reasoning and reinforcement learning with verifiable rewards to enhance accuracy on complex problems. Hector Liu, director of MBZUAI's institute of foundation models, said the project is treated "more like a system than just a model." Unlike typical open-source releases, K2 Think is being actively deployed and iteratively improved, reflecting a long-term development strategy. K2 Think is built on Alibaba's Qwen 2.5 large language model and runs on hardware from Cerebras. MBZUAI claims it is among the fastest and most efficient reasoning systems, processing around 2,000 tokens (about 1,500 words) per second. This balance of speed and accuracy is intended to give it a competitive edge against larger, more resource-intensive models. Like DeepSeek's R1 model, K2 Think is open source. Its training data and weights are publicly available, allowing researchers worldwide to study, reproduce, and extend its reasoning methods. MBZUAI says this transparency will encourage collaboration and accelerate progress in reasoning models. The university calls K2 Think "a defining moment for AI in the UAE," emphasizing how open innovation and public-private partnerships can position Abu Dhabi as a global AI hub. By investing in efficient and transparent AI systems, the UAE aims to influence how reasoning models evolve beyond scale-driven approaches.

MBZUAI, G42 launch open-source AI reasoning system K2 Think

Image: Getty Images/ For illustrative purposes The Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) and Abu Dhabi-based G42 launched K2 Think, an open-source system for advanced AI reasoning that delivers performance on par with the most powerful reasoning models at a fraction of their size. With 32 billion parameters, K2 Think outperforms flagship reasoning models nearly 20 times larger, according to MBZUAI's Institute of Foundation Models. The system ranks among the top reasoning models globally, leading all open-source models in math benchmarks including AIME '24/'25, HMMT '25 and OMNI-Math-HARD. What K2 Think offers Built on six pillars of innovation, K2 Think employs chain-of-thought supervised fine-tuning, reinforcement learning with verifiable rewards, agentic planning, and test-time scaling. It will also be available on Cerebras' wafer-scale compute platform, optimised for speculative decoding at 2,000 tokens per second. Khaldoon Khalifa Al Mubarak, chairman of MBZUAI's Board of Trustees and Member of the Artificial Intelligence and Advanced Technology Council (AIATC), said: "The new global benchmark set by K2 Think underscores the pioneering excellence of MBZUAI's Institute of Foundation Models initiative. It is also an example of the UAE's commitment to building advanced systems that are developed by our institutions and shared with the world." Peng Xiao, MBZUAI board member and group CEO of G42, said: "K2 Think has shifted the AI reasoning paradigm from 'bigger is better' to 'smarter is better'... By proving that smaller, more resourceful models can rival the largest reasoning systems, this milestone marks the beginning of the next wave of AI innovation." Unlike most models that only release weights, K2 Think is fully open source, including training data, weights, and deployment software. Professor Eric Xing, MBZUAI president, said the system represents "a new era of cost-effective, reproducible and accountable AI." MBZUAI's open-source family of models K2 Think extends MBZUAI's open-source family of models including Jais (Arabic), NANDA (Hindi) and SHERKALA (Kazakh), and builds on K2-65B, the world's first fully reproducible open-source foundation model released in 2024. K2 Think is available at K2Think.Ai/K2Think and on Hugging Face. Read: MBZUAI's Rawdha Almeraikhi on how Emirati women are driving innovation