Moonshot AI
Moonshot AI develops frontier foundation models and AI products under the Kimi brand, with a focus on converting energy into intelligence and making advanced intelligence broadly accessible through open models. Its products target consumer assistance, knowledge work, coding, research, and enterprise workflows.
At a Glance
- Consumers
- Software developers and coding teams
- Enterprises and knowledge-work teams
- Investment banks and securities firms
- +4 more
AI Tools by Moonshot AI
(6)
Kimi API Platform
LLM API Platform for Developers
Kimi Vendor Verifier
Kimi Model Provider Verification
Kimi Code CLI
Terminal AI Coding Agent
Kimi
Agentic AI Workspace Platform
Kimi Claw
Cloud Hosted AI Agent Platform
Kimi Code
AI Coding Assistant
Discussions
The Price of Intelligence Is Collapsing. OpenAI and Anthropic Should Be Worried.
The conventional wisdom in enterprise AI has been that you pay a premium for the best model, and the best models come from OpenAI and Anthropic. Kimi K3 does not disprove that entirely, but it puts a hard expiration date on how long that premium can hold. This is not primarily a story about benchmar…
Latest News
Moonshot announced Kimi K3.1 Pro, a 3T-class flagship with 1M-token context and native visual understanding.
Bloomberg reported Moonshot completed a final private financing at about a $50 billion valuation and was targeting a Q1 2027 Hong Kong IPO.
Moonshot launched Kimi's financial-industry AI solution with 10+ data sources, nine financial skills, and five compliance/security measures.
Bloomberg reported Moonshot ARR topped $1 billion in August and that the company was targeting $2 billion in annualized revenue by year-end.
Products & Services
Consumer AI assistant available on the web and mobile, built around long-context conversation, search, research, document analysis, and agentic task execution.
Developer platform providing usage-based access to Kimi models through Chat Completions, Responses, and Anthropic-compatible Messages APIs.
Reasoning model based on reinforcement learning, designed for long-context math, coding, and multi-step problem solving.
Open-weight trillion-parameter MoE model for coding, reasoning, and agentic workflows.
Market Position
Moonshot positions Kimi as an open-model, long-context and increasingly agentic alternative to proprietary frontier systems. Its models compete with OpenAI's GPT/ChatGPT, Anthropic's Claude, Google's Gemini, and Chinese systems including ByteDance Doubao, Alibaba Qwen, Zhipu's Z.ai, and DeepSeek; differentiation centers on open weights, very large context windows, multimodal and coding capability, agent swarms, and comparatively low inference cost.
Leadership
Founders
Yang Zhilin
Computer-science researcher with a PhD from Carnegie Mellon University; previously worked with Google Brain and Meta AI and co-founded Recurrent AI. He is Moonshot AI's CEO.
Zhou Xinyu
Tsinghua University computer-science alumnus who previously worked at Hulu, Tencent, and Megvii on research and deployment of deep-neural-network systems.
Wu Yuxin
Computer-vision and deep-learning researcher; previously worked at Google Brain on foundation models and at Facebook AI Research on computer vision, where he created or contributed to Detectron2.
Executive Team
Yang Zhilin
Co-founder and Chief Executive Officer
CMU PhD and former Google Brain and Meta AI researcher; previously co-founded Recurrent AI.
Zhang Yutao
Co-founder and Chief Technology Officer
Tsinghua-trained AI researcher and co-founder of Recurrent AI; associated with Moonshot's model research and technology development.
Board of Directors
Founding Story
Founded in Beijing in spring 2023 by three Tsinghua-linked researchers, Moonshot AI was driven by a goal of building AGI and exploring the limits of intelligence. The company's name and 'far side of the moon' metaphor express its ambition to pursue what lies beyond the edge of current human and technological understanding; its early product strategy emphasized long-context foundation models and open models.
Business Model
Revenue Model
Usage-based API inference billed per input/output token, plus paid consumer Kimi memberships and enterprise products such as Kimi Work and the financial-industry solution.
Pricing Tiers
Prompt-cache hit $0.20/MTok; 5-minute cache writes $2.00/MTok; 1-hour cache writes $4.00/MTok.
High-speed K3.1 Pro variant; cache hit $0.20/MTok.
Pay-as-you-go API; cache hit $0.30/MTok.
CNBC reported this subscription range for Kimi, with tier-dependent access to data and capabilities.
Target Markets
- Consumers
- Software developers and coding teams
- Enterprises and knowledge-work teams
- Investment banks and securities firms
- Commercial banks and insurance companies
- Venture capital and private-equity funds
- Consumer question answering and personal productivity
- Deep research and long-document synthesis
- Autonomous coding, debugging, refactoring, and repository-scale software development
- Legal and intellectual-property document review
- Financial research, valuation, portfolio review, and investment-banking workflows
- Scientific research and AI-for-science workflows
- CICC
- Sequoia China / HongShan
- FunPlus
- Tencent CodeBuddy