Another Chinese AI model is turning heads
Alibaba-backed startup Moonshot released on late Friday night its Kimi K2 model: a low-cost, open source large language model — the two factors that underpinned China-based DeepSeek's industry disruption in January. Open-source technology provides source code access for free, an approach that few U.S. tech giants have taken, other than Meta and Google to some extent.
Coincidentally, OpenAI CEO Sam Altman announced early Saturday that there would be an indefinite delay of its first open-source model yet again due to safety concerns. OpenAI did not immediately respond to a CNBC request for comment on Kimi K2.
One of Kimi K2′s strengths is in writing computer code for applications, an area in which businesses see potential to reduce or replace staff with generative AI. OpenAI's U.S. rival Anthropic focused on coding with its Claude Opus 4 model released in late May.
In its release announcement on social media platforms X and GitHub, Moonshot claimed Kimi K2 surpassed Claude Opus 4 on two benchmarks, and had better overall performance than OpenAI's coding-focused GPT-4.1 model, based on several industry metrics.
'No doubt [Kimi K2 is] a globally competitive model, and it's open sourced,' Wei Sun, principal analyst in artificial intelligence at Counterpoint, said in an email Monday.
Cheaper option
'On top of that, it has lower token costs, making it attractive for large-scale or budget-sensitive deployments,' she said.
The new K2 model is available via Kimi's app and browser interface for free unlike ChatGPT or Claude, which charge monthly subscriptions for their latest AI models.
Kimi is also only charging 15 cents for every 1 million input tokens, and $2.50 per 1 million output tokens, according to its website. Tokens are a way of measuring data for AI model processing.
In contrast, Claude Opus 4 charges 100 times more for input — $15 per million tokens — and 30 times more for output — $75 per million tokens. Meanwhile, for every one million tokens, GPT-4.1 charges $2 for input and $8 for output.
Moonshot AI said on GitHub that developers can use K2 however they wish, with the only requirement that they display 'Kimi K2' on the user interface if the commercial product or service has more than 100 million monthly active users, or makes the equivalent of $20 million in monthly revenue.
Hot AI market
Initial reviews of K2 on both English and Chinese social media have largely been positive, although there are some reports of hallucinations, a prevalent issue in generative AI, in which the models make up information.
Still, K2 is 'the first model I feel comfortable using in production since Claude 3.5 Sonnet,' Pietro Schirano, founder of startup MagicPath that offers AI tools for design, said in a post on X.
Moonshot has open sourced some of its prior AI models. The company's chatbot surged in popularity early last year as China's alternative to ChatGPT, which isn't officially available in the country. But similar chatbots from ByteDance and Tencent have since crowded the market, while tech giant Baidu has revamped its core search engine with AI tools.
Kimi's latest AI release comes as investors eye Chinese alternatives to U.S. tech in the global AI competition.
Still, despite the excitement about DeepSeek, the privately-held company has yet to announce a major upgrade to its R1 and V3 model. Meanwhile, Manus AI, a Chinese startup that emerged earlier this year as another DeepSeek-type upstart, has relocated its headquarters to Singapore.
Over in the U.S., OpenAI also has yet to reveal GPT-5.
Work on GPT-5 may be taking up engineering resources, preventing OpenAI from progressing on its open-source model, Counterpoint's Sun said, adding that it's challenging to release a powerful open-source model without undermining the competitive advantage of a proprietary model.
Grok 4 competitor
Kimi K2 is not the company's only recent release. Moonshot launched a Kimi research model last month and claimed it matched Google's Gemini Deep Research 's 26.9 score and beat OpenAI's version on a benchmark called 'Humanity's Last Exam.'
The Kimi research model even got a mention last week during Elon Musk's xAI release of Grok 4 — which scored 25.4 on its own on the 'Humanity's Last Exam' benchmark, but attained a 44.4 score when allowed to use a variety of AI tools and web search.
'Kimi-Researcher represents a paradigm shift in agentic AI,' said Winston Ma, adjunct professor at NYU School of Law. He was referring to AI's capability of simultaneously making several decisions on its own to complete a complex task.
'Instead of merely generating fluent responses, it demonstrates autonomous reasoning at an expert level — the kind of complex cognitive work previously missing from LLMs,' Ma said. He is also author of 'The Digital War: How China's Tech Power Shapes the Future of AI, Blockchain and Cyberspace.'
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
NBC News
6 hours ago
- NBC News
Anthropic, Google, OpenAI and xAI granted up to $200 million for AI work from Defense Department
The U.S. Department of Defense on Monday said it's granting contract awards of up to $200 million for artificial intelligence development at Anthropic, Google, OpenAI and xAI. The DoD's Chief Digital and Artificial Intelligence Office said the awards will help the agency accelerate its adoption of 'advanced AI capabilities to address critical national security challenges.' The companies will work to develop AI agents across several mission areas at the agency. 'The adoption of AI is transforming the Department's ability to support our warfighters and maintain strategic advantage over our adversaries,' Doug Matty, the DoD's chief digital and AI officer, said in a release. Elon Musk's xAI also announced Grok for Government on Monday, which is a suite of products that make the company's models available to U.S. government customers. The products are available through the General Services Administration (GSA) schedule, which allows federal government departments, agencies, or offices to purchase them, according to a post on X. OpenAI was previously awarded a year-long $200 million contract from the DoD in 2024, shortly after it said it would collaborate with defense technology startup Anduril to deploy advanced AI systems for 'national security missions.' In June, the company launched OpenAI for Government for U.S. federal, state, and local government workers.
Finextra
8 hours ago
- Finextra
The Future of AI: Opportunities and Risks in the Next Decade: By Erica Andersen
Predicting where artificial intelligence will be in 5-10 years is challenging given the rapid pace of change, but one thing is certain: AI will have a profound impact on how we work and live. Having observed senior managers grapple with AI implementation at recent industry summits, it's clear that while leadership understands change is coming, most organizations are still "nibbling around the corners" rather than embracing transformative applications. The Current Implementation Challenge Organizations face a fundamental concern in AI adoption. Many want to maintain data silos and restrict AI access to information within their own walls, creating a recurring question: what information should AI systems have access to? This cautious approach, while understandable from a security perspective, may limit AI's transformative potential. The desire to maintain data silos in the context of AI adoption is a complex apprehension issue driven by a combination of factors, including data security, competitive advantage, regulatory compliance, technical challenges, and organizational culture. While data silos can offer benefits in terms of control and security, they can also hinder innovation and limit the potential of AI. Organizations must carefully weigh these competing considerations when developing their AI strategies. Slow adoption isn't entirely negative. Senior executives often lack deep technical understanding, and this delay has given us time to better understand how to use large language models (LLMs) effectively. We're still in the experimental phase, much like the early days of steam engines when we could make them work without fully understanding the underlying principles. Debunking the Job Replacement Myth The widespread claim that AI will eliminate masses of jobs deserves scrutiny. While AI will likely reduce certain types of work, the jobs most often cited as at risk—programing and data analysis—may actually become more valuable. History shows us that as workers become more efficient, organizations don't run out of tasks; they find new ways to leverage that efficiency. The charts showing AI-related layoffs are misleading. Most of these reductions began in 2022, well before ChatGPT gained prominence. The real culprit was likely Section 174 of the U.S. tax code, which required companies to capitalize all R&D salaries—a change that forced many organizations to trim their technical workforce to manage new tax obligations. Learning from History: The Steam Engine Parallel The AI revolution mirrors the steam engine transformation more than any other technological shift. Both technologies emerged before we fully understood their mechanics. Just as thermodynamics was developed by Carnot to improve steam engine efficiency during the Napoleonic Wars, while AI advancements are rapid, the industry is still groping to understand optimal AI architectures—from embedding space sizes to attention mechanisms. Most importantly, the creators of steam engines weren't the ultimate commercial winners. Similarly, AI engine developers may not capture the greatest value. The lack of strong monopoly effects, combined with inevitable competition (particularly from China's open-source initiatives), suggests that AI infrastructure providers face uncertain long-term prospects. Where the Real Opportunities Lie The biggest winners will likely be organizations that successfully integrate AI into their operations, gaining a significant competitive advantage, just as textile manufacturers who adopted steam engines transformed their industries took advantage of that pivotal moment in the Industrial Revolution. The Industrial Revolution saw a pivotal shift driven by visionary textile manufacturers like Richard Arkwright and Samuel Greg. Recognizing the transformative power of steam, pioneers like James Watt and Matthew Boulton invested in its potential, building steam engines to revolutionize production. They understood that steam power could increase output, lower labor costs, and streamline manufacturing processes, effectively laying the groundwork for modern industry. While these innovators focused on the application of steam, Francis Cabot Lowell is credited with bringing the Industrial Revolution to the United States, though his reliance on steam power at that early stage is uncertain. Examining the potential of AI today reveals particular promise across several sectors: Knowledge Management – a New Revolution in the Making? AI will fundamentally change how organizations capture and apply tribal knowledge. Rather than managing data, AI will revolutionize knowledge management—how we organize, share, and gain insights from information. This represents a true "Knowledge Revolution 2.0." Enhanced Professional Services Legal professionals will benefit from advanced search capabilities through RAG (Retrieval-Augmented Generation) systems and scenario planning tools. Education will see improvements in lesson planning, grading, and personalized student feedback. These aren't job replacements but significant productivity enhancements. Operational Excellence Vision applications, already advanced before ChatGPT's emergence, offer tremendous potential for monitoring and improving operations. Companies are far behind in recognizing these capabilities. AI-powered project management tools will provide unprecedented visibility into progress and problems, potentially preventing costly delays like those seen in major infrastructure projects. Sales and Customer Relations Customer relationship management (CRM) will evolve from clunky data entry systems to comprehensive, automated platforms. These platforms will automatically monitor interactions (calls, emails, meetings) and consolidate data from diverse sources, creating a unified view of each customer and partner. This transformation will free sales professionals to focus on relationship building rather than administrative tasks. Content and Creative Industries Rather than eliminating creative jobs, AI tools may democratize content creation and drive down costs. The current £5,000 price tag a website expert needs for basic website development reflects tool limitations rather than fair market value. Improved automation could increase both the supply of and demand for creative services. The Infrastructure Question The "picks and shovels" strategy—investing in AI infrastructure providers—faces significant risks. Efficiency improvements, as demonstrated by DeepSeek's cost-effective model training, could rapidly devalue existing hardware investments. The key insight, seemingly lost on some major players, is that moving data efficiently through systems matters more than raw computational power. Moreover, geopolitical tensions may accelerate the development of alternative infrastructure providers, particularly in China, potentially fragmenting what some assumed would be a stable monopoly market. The Path Forward Over the next decade, large enterprises must embrace AI or risk being overtaken by competitors who successfully integrate these capabilities. I believe the most significant transformation won't be in data processing but in knowledge management—how organizations make decisions, solve problems, and leverage collective intelligence. Leadership teams currently focused on risk, governance, and security—while important—are missing their critical role as champions of revolutionary technology. They need to shift from defensive positioning to actively exploring where and how AI can maximize their organization's potential, especially within the tribal knowledge sphere. The organizations that thrive will be those that move beyond tentative experimentation to thoughtful, comprehensive AI integration. The question isn't whether AI will transform business operations, but whether your company will lead or follow in this transformation. As we enter this new era, success will depend less on access to the most advanced AI models and more on the wisdom to apply these tools effectively to real business challenges. The revolution is underway—the winners will be those who recognize that we're not just in an AI revolution, but in Knowledge Revolution 2.0. Written by Oliver King-Smith, founder and CEO, smartR AI
Reuters
9 hours ago
- Reuters
US defense department awards contracts to Google, Musk's xAI
July 14 (Reuters) - OpenAI, Alphabet's Google (GOOGL.O), opens new tab, Anthropic and Elon Musk's AI firm xAI have won contracts aimed at scaling up adoption of advanced AI capabilities in the U.S. Department of Defense, the Chief Digital and Artificial Intelligence Office said on Monday. Each of the contracts has a $200 million ceiling and will enable the DoD to develop agentic AI workflow and use them to address critical national security challenges, the office said. "Establishing these partnerships will broaden DoD use of and experience in frontier AI capabilities and increase the ability of these companies to understand and address critical national security needs," it said. The Pentagon last month announced OpenAI was awarded a $200 million contract, saying the ChatGPT maker would "develop prototype frontier AI capabilities to address critical national security challenges in both warfighting and enterprise domains." The White House's Office of Management and Budget released new guidance in April, directing federal agencies to ensure that the government and "the public benefit from a competitive American AI marketplace.
