kimi

The Dark Side of the Moon (Kimi) has been renamed Beijing Dark Side of the Moon Technology Co., Ltd., and the company type has changed from a limited liability company to a non-listed joint-stock company. Yang Zhilin serves as the chairman and manager, Zhang Yutong has been added as a director, and Song Sijia has been added as the financial officer.This change is commonly referred to as a stock reform. The China Securities Regulatory Commission stipulates that if a domestic company directly lists overseas with a domestic entity, the issuer must be a joint-stock company.According to previous reports, the Dark Side of the Moon has issued a shareholder resolution for listing in Hong Kong, with the earliest possible IPO within six months, and has begun dismantling the red-chip structure. The company has not officially announced the listing time.

According to news from Alibaba Cloud, the Alibaba Cloud Zhenwu M890 super node instance has been Day0 adapted to Kimi K3. Both parties will further develop domestic computing power cooperation, and the Qianwen AI platform and Alibaba Cloud Bailian will also provide the model API for Kimi K3.

The Kimi.ai team has opened the Kimi K3 model weights and technical report while gradually open-sourcing multiple underlying infrastructures: including the high-performance communication library MoonEP for distributed MoE training, the distributed environment system AgentENV for large-scale agent workflows (in collaboration with kvcache-ai), and the Kimi Delta Attention high-performance kernel FlashKDA based on CUTLASS.The official statement claims that these components can reduce communication and inference overhead in large-scale MoE and agent reinforcement learning training, and can be used as a plug-and-play backend for flash-linear-attention.

The joint assessment by the UK Institute for AI Safety and the US Center for AI Standards and Innovation shows that Moonshot AI's Kimi K3 can assist in the development of exploit vulnerabilities and simulate cyber attacks without substantial hindrance, but its overall cybersecurity capabilities still lag significantly behind leading models in the United States.

Moonshot AI's latest flagship open-source model Kimi K3 is officially launched on B.AI Web Chat. Previously, B.AI had already completed the Kimi K3 API integration, and this Web Chat launch further lowers the usage threshold. Users can now directly experience the powerful capabilities of this world's first open-source 3T level model in long-range programming, complex reasoning, and deep knowledge work on the web.

NVIDIA CEO Jensen Huang stated in an exclusive interview with Axios on Tuesday that American companies "absolutely" should be allowed to use Chinese open-source AI models, directly challenging the Trump administration and some American AI labs' blockade policies.Huang believes that the market's panic over Kimi K3 is a misreading, similar to the sell-off triggered by DeepSeek in early 2025: cheaper open-source models will expand the AI audience and increase, rather than decrease, the demand for chips, data centers, and computing power. "Free AI is good for hardware, good for chips, good for data centers." He also refuted the notion that open-source models pose security risks, claiming that open-source is actually safer because external researchers can examine the models, expose vulnerabilities, and build defenses, while calling for Anthropic to open its Claude Mythos model to "everyone."Huang rejected the narrative that "China will defeat American companies," arguing that the AI race has no finish line and that China and the U.S. will coexist in the long term. Hours after the interview, U.S. Treasury Secretary Bessent stated that the government is reviewing whether Chinese AI models are stealing intellectual property and considering sanctions. Huang responded that knowledge distillation is the foundation of intelligence and that accountability should be directed at violations rather than the models themselves.

The Chinese AI large model company Dark Side of the Moon is accelerating its capitalization process, planning to start the final round of pre-IPO financing negotiations in August, with a target pre-investment valuation of $50 billion. Reports indicate that Dark Side of the Moon expects to complete a round of financing soon, corresponding to a valuation of approximately $31.5 billion. After completing this round of financing, the company will immediately engage in discussions for the next round of financing with potential investors, which may become its last private placement before entering the Hong Kong capital market.Market participants believe that if the financing plan proceeds smoothly, Dark Side of the Moon could initiate the process for a Hong Kong stock listing as early as this year. One of the key driving factors for this valuation increase is the company's recent release of the new generation large model Kimi K3. It is reported that Kimi K3 has a parameter scale of 28 trillion, supporting a context window of millions of tokens, making it one of the largest open-source models in the world by parameter scale. After the model was released, it quickly attracted market attention, prompting the company to adjust its new user subscription arrangements to ensure the user experience for existing customers.As the competition in generative AI enters a phase of comparison in computing power, user scale, and commercialization capabilities, large model companies are accelerating their search for capital support. Dark Side of the Moon's financing plan also reflects the market's ongoing interest in leading AI foundational model companies. However, the $50 billion valuation target also implies higher commercialization requirements. In the future, investors will focus on Dark Side of the Moon's revenue growth, control of computing costs, and whether the Kimi series models can form a stable business model. If the financing and listing plans proceed smoothly, Dark Side of the Moon may become an important case of capitalization for Chinese AI large model companies.

Fireworks AI released an evaluation report comparing the open-source model Kimi K3 and the closed-source model Fable 5 across approximately 1030 agent tasks (covering scenarios such as SWE, terminal operations, algorithms, multilingual coding, and legal matters). The results showed that the accuracy rates for the two models on the SWE benchmark were 92.4% and 92.6%, respectively, indicating similar overall performance, but each model has its strengths in specific areas: K3 leads in symbolic mathematics and development tools, while Fable excels in web and data visualization; in long-cycle terminal tasks, K3 independently tackled 11 tasks that Fable could not solve.The report pointed out that adopting a task-level routing strategy can dynamically allocate between the two, achieving an accuracy rate of 93%, surpassing either single model. Additionally, due to K3's significant cost advantage on the Fireworks platform, costs for long-cycle agent tasks can be up to 50 times lower than Fable. The report recommends using the open-source model as the default option and continuously learning the optimal match between tasks and models through routers to balance quality and cost.

According to market news, Goldman Sachs reported on July 18 that the Dark Side of the Moon released the Kimi K3 model, with 28 trillion parameters, surpassing Claude Fable 5 and GPT-5.6 Sol on the Arena.ai programming leaderboard, reaching a new high for Chinese models with an API pricing of $2.3 per million tokens. Two days after the release, Zhiyu fell by 28%, MiniMax dropped by 16%, the Nasdaq 100 index futures fell by more than 1.8%, and the Philadelphia Semiconductor Index has cumulatively dropped by more than 18% from its peak.Goldman Sachs believes that Kimi K3 marks a turning point: a Chinese laboratory that cannot match the largest pre-trained computing power in the West has rapidly narrowed the gap with top American models through architectural innovation and reinforcement learning, proving that "scaling" is no longer the only winning path. Goldman Sachs warns that the "era of computing power expansion" may be coming to an end, and the investment logic for AI infrastructure built around "the more computing power, the better" needs to be rewritten. Goldman Sachs maintains a buy rating on MiniMax and a neutral rating on Zhiyu. Future attention will focus on the intensive release of 20-50 trillion parameter models such as Zhiyu GLM, Alibaba Qwen, and MiniMax M3 Pro.

According to official news, Kimi stated that in the past 48 hours, the demand for Kimi K3 has exceeded expectations, and GPU resources are nearing the current capacity limit. To ensure the experience of existing subscription users, the platform has temporarily suspended new subscriptions, and current subscribed users will not be affected. Kimi mentioned that it is working to expand computing power as soon as possible and will gradually open new subscription slots in batches.Meanwhile, the platform will split the membership system into two types of plans in the future: one for Kimi members on the Kimi web portal, App, and Work, and the other for Kimi Code members focused on programming workflows, to more accurately match computing resources and maintain service stability.

B.AI announced that the Kimi K3 model is officially launched on its API platform, seamlessly integrated after the release by @Kimi_Moonshot. As the world's first open-source 3T level (28 trillion parameters) model, Kimi K3 has made significant breakthroughs in long-range programming, complex reasoning, and knowledge work, and natively supports visual capabilities and 1M ultra-long context. As a model aggregation service platform, B.AI has prioritized opening its API interface, allowing developers to directly experience the capabilities of the new generation model through B.AI. B.AI will continue to provide efficient and stable interface services to promote the rapid implementation of cutting-edge AI technology and empower developers' innovation.

According to Xinhua News Agency, as the 2026 World Artificial Intelligence Conference and the High-Level Meeting on Global Governance of Artificial Intelligence are about to be held, the new generation model Kimi K3 was released on the 16th, with a parameter scale of 28 trillion. This is currently the largest open-source model in the world by parameters, marking a new step in the development of artificial intelligence models in our country.Beijing Moon's Dark Side Technology Co., Ltd., which released the model, introduced that Kimi K3 natively supports visual understanding, has a context window of 1 million tokens, and is optimized for complex task scenarios such as software engineering, knowledge work, deep research, and multimodal understanding, further enhancing the processing capability of large models for complex tasks. It is understood that in evaluations, Kimi K3's comprehensive intelligence level is close to that of the world's leading closed-source models. A relevant person in charge of Moon's Dark Side stated that in the process of training the new generation foundational model Kimi K3, they used a self-developed underlying model architecture and accumulated a complete set of scientific model training methods.

According to the Financial Times, informed sources reveal that the Chinese AI unicorn company Moonshot AI plans to release a new large language model, Kimi K3, in the near future. This model has between 20 trillion to 30 trillion parameters, making it the largest AI model in China by parameter scale, and its performance is expected to surpass the flagship model Claude Opus 4.8 from Anthropic in mainstream benchmark tests (industry speculation suggests its parameter count is around 15 trillion to 20 trillion).Unlike the currently mainstream closed-source and expensive cutting-edge large models in the United States, Kimi K3 will be available as an open-weight model for users to download and modify for free, which may create competitive pressure for leading American labs like OpenAI and Anthropic. Currently, due to the rising service fees for large models in the U.S. (for example, Anthropic has announced a 50% price increase for Opus 4.8 in September), some overseas companies have begun to shift towards using more cost-effective Chinese open-source models.In terms of the capital market, informed sources indicate that Moonshot AI is preparing for a new round of financing, with the latest valuation expected to reach approximately $31.5 billion. Meanwhile, the valuations of other AI giants in China and the U.S. are also rising; DeepSeek is starting a new round of financing with an estimated valuation of about $71 billion, while Anthropic and OpenAI have reached valuations of $965 billion and $852 billion, respectively, in their latest round of financing. In response to the aforementioned release and financing rumors, Moonshot AI has currently declined to comment.

According to GMGN market data, the market capitalization of the BSC chain meme coin Hakimi briefly surpassed $17 million, then fell back to around $16 million, with a 24-hour increase of over 50%.The price of meme coins is highly volatile, mainly driven by market sentiment and conceptual speculation, lacking clear practical application scenarios. Investors should be aware of the associated risks.

According to the Science and Technology Innovation Board Daily, the recent round of financing for Moon's Dark Side Kimi, with a valuation of 20 billion USD, has been completed, and a new round of financing has been initiated, with a pre-financing valuation rising to 31.5 billion USD. According to sources close to Kimi, the company disclosed its latest revenue data during this round of financing communication: ARR (Annual Recurring Revenue) has surpassed 300 million USD. The revenue growth for Kimi in this round mainly comes from increased developer usage and API revenue driven by model iterations. Currently, API revenue accounts for over 70% of Kimi's total revenue and continues to rise. Kimi's revenue curve is beginning to show characteristics of the early commercialization stage of Anthropic: increased developer usage, a higher proportion of API revenue, growth in overseas paying users, and a price system uplift driven by model capability iterations.

According to a report by 21 Finance, Huang Zhenxin, the head of Kimi B-end at Moonshot AI, stated in a recent communication meeting that there is indeed a bubble in the current AI industry, but the fundamentals are very solid. Enterprises can now clearly calculate the return on investment (ROI), and the substantial transformation in productivity brought by AI has already occurred.Regarding the recent phenomenon of widespread price increases among model vendors, Huang Zhenxin pointed out that the core reason lies in the rising global computing power costs, and chip production capacity cannot meet the explosive growth in Token demand. He emphasized that evaluating the cost-performance ratio of models should not only look at the unit price of input and output but should also focus on the Cache hit rate. It is reported that Kimi's original factory Cache hit rate has reached over 90%, significantly reducing actual computing costs.In addition, Huang Zhenxin revealed that Kimi will continue to challenge innovations in underlying architecture to sustain the Scaling Law, and its Muon optimizer, which has been validated on a large scale, is now widely adopted by several mainstream large models in the industry. Regarding the "last mile" of enterprise AI implementation, he believes that as the foundational capabilities of models continue to strengthen, the technical paradigms at the application layer will also continue to simplify.

According to GMGN monitoring, the Meme coin "Hakimi" on the BSC chain surged significantly yesterday, with its market cap briefly exceeding $22 million. Due to the continued decline of BTC, the token faced pressure and has now reported approximately $18.4 million, with a 24-hour increase still holding at 23%. During the same period, the trading volume reached $2.9 million, making it the Meme coin with the largest trading volume on the BSC chain for the day.ChainCatcher reminds users that Meme coin trading is highly volatile, largely relying on market sentiment and speculative hype, and has no actual value or use case. Investors should be aware of the risks.

According to LatePost, Kimi (Dark Side of the Moon) is about to complete a new round of $2 billion financing, with a post-investment valuation exceeding $20 billion. This round of financing is led by Meituan Longzhu, with participation from China Mobile, CPE (CITIC Industrial Fund), and others, with Longzhu alone contributing over $200 million. In January and February of this year, Kimi had just completed three rounds of financing, raising $500 million, $700 million, and $700 million respectively. Including this latest financing, Kimi has raised over $3.9 billion in less than six months, with the latest valuation more than quadrupling from about $4.3 billion in November of last year.So far, Kimi's total financing amount has exceeded 37.6 billion RMB, making it the company with the highest cumulative financing among large model startups. Including IPO fundraising, MiniMax has raised approximately 15 billion RMB, and Zhipu around 13 billion RMB. As of today’s midday market close, MiniMax's market capitalization is approximately 210 billion RMB, while Zhipu is about 347 billion RMB.