TL;DR
Get hardware and tech essentials delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
Moonshot AI released Kimi K3, a 2.8 trillion parameter model priced at $3 per million input tokens, matching Western mid-tier models in cost. This challenges China’s previous cost advantage and signals a shift in AI capability competition.
Moonshot AI announced the release of Kimi K3 on July 16, 2026, a new AI model with 2.8 trillion parameters priced at $3 per million input tokens. This makes it the most expensive Chinese model to date and places it on par with Western mid-tier models, marking a major shift in China’s AI market and competitive stance.
Moonshot’s Kimi K3 is a 2.8 trillion parameter model, surpassing previous Chinese models in scale. It is priced at $3 per million input tokens and $15 per million output tokens, matching the standard rate of Western models like Claude Sonnet 5. This pricing indicates a move away from the cheap, accessible Chinese AI models that dominated the past two years, signaling increased confidence in Chinese capabilities.
The model is built using a sparse Mixture-of-Experts architecture, with 16 of 896 experts active per token, and features a 1,048,576-token context window, supporting native text, image, and video input. It is currently available via API, in the Kimi app, and the Playground platform. Independent benchmarks show Kimi K3 performing well against global models, ranking fourth in some tests and close to the top in others, with a score of 57.1 on the Artificial Analysis Intelligence Index v4.1, just 2.8 points below the leading models.
While Moonshot has promised to release the model weights by July 27, the current availability is limited to hosted API access, and the active parameter count remains undisclosed. The pricing and capability level mark a significant shift, challenging the narrative that Chinese AI is only cost-effective due to efficiency constraints.
Impact of Kimi K3 on Global AI Competition
The release of Kimi K3 at Western mid-tier prices, combined with its high capability, signals a major shift in China’s AI industry. It undermines the previous narrative that Chinese models could only compete on cost, suggesting that domestic labs are now capable of producing large, high-performance models without relying solely on efficiency or export restrictions. This development could accelerate China’s entry into the top tier of AI research and commercial deployment, reshaping global market dynamics and competitive strategies.
As an affiliate, we earn on qualifying purchases.
Background on Chinese AI Development and Market Shifts
Over the past two years, Chinese AI labs have focused on developing cost-effective, efficient models due to export controls and resource limitations. The common belief was that Chinese models would remain behind Western counterparts in scale and capability, but at a lower cost. Moonshot AI’s previous models, such as K2, were around 1 trillion parameters, and the industry largely viewed Chinese AI as an affordable alternative rather than a frontier competitor.
The recent launch of Kimi K3 challenges this view, as it is the largest open-weight model announced globally and is priced at the same level as Western mid-tier models. This suggests a possible acceleration in Chinese AI research and a shift in the competitive landscape, with domestic labs potentially overcoming previous constraints on compute and scale.
Analysts expected China to reach this level of capability by early 2027; instead, it has achieved it six months early, raising questions about the effectiveness of export controls and the true state of technological independence.
“Our most capable model to date, with 2.8 trillion parameters, demonstrates China’s rapid advancement in AI capabilities.”
— Yutong Zhang, President of Moonshot AI
large-scale AI model for developers
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unconfirmed Aspects of Kimi K3’s Active Parameters
While the total parameter count is confirmed at 2.8 trillion, the number of active parameters during training remains undisclosed. This gap affects interpretations of the model’s compute efficiency and true scale, and whether the large parameter count translates into comparable performance and resource requirements.
Additionally, the long-term performance and real-world deployment efficacy of Kimi K3 are still to be evaluated as independent benchmarks and user feedback emerge.
AI model training and deployment tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps for Kimi K3 and Chinese AI Progress
Moonshot AI plans to release the model weights by July 27, which will enable broader independent evaluation of Kimi K3’s capabilities. The company will also continue to refine the model and expand its deployment through APIs and applications.
Industry analysts will monitor whether this development prompts other Chinese labs to accelerate their own large-scale model projects, potentially leading to a new phase of competitive parity or even dominance in AI capabilities. The impact on export controls and international policy remains uncertain and will be closely watched.
AI model performance benchmarking software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What makes Kimi K3 different from previous Chinese models?
Kimi K3 features 2.8 trillion parameters, making it the largest open-weight Chinese model, and is priced at Western mid-tier levels, signaling a shift from cost-focused to capability-focused competition.
Why is the pricing of Kimi K3 significant?
Pricing at $3 per million input tokens aligns it with Western models like Claude Sonnet 5, indicating Chinese labs are confident in their model’s capabilities and are no longer competing solely on cost.
What are the implications for global AI competition?
This development suggests China’s AI industry is reaching the frontier faster than expected, potentially challenging Western dominance and altering market and policy dynamics.
When will independent evaluations of Kimi K3’s performance be available?
Independent benchmarks are expected after the release of the model weights, anticipated by July 27, 2026.
Does this mean export restrictions are ineffective?
The existence of such a large-scale, high-capability model raises questions about the effectiveness of export controls, but the full implications are still under analysis.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
