According to the Financial Times, Moonshot AI’s upcoming Kimi K3, with 2‑3 trillion parameters, will become China’s largest open‑weight model and aim to close the gap with Anthropic’s Opus 4.8, while the firm seeks a new $31.5 billion valuation.

Key Takeaways

  • Kimi K3 will house 2‑3 trillion parameters
  • China’s largest open‑weight AI model
  • Targets parity with Anthropic’s Opus 4.8
  • Moonshot AI raises funds at a $31.5 billion valuation

Moonshot AI, one of China’s leading artificial‑intelligence labs, has announced the third generation of its Kimi series – the Kimi K3. Citing anonymous sources, the Financial Times confirms that the model will contain between 2 and 3 trillion parameters, making it the largest open‑weight AI system to emerge from China.

Success of the Previous Generation

The Kimi K2 models have already earned a solid reputation in the open‑source AI market, ranking highly on benchmark suites and delivering capabilities that sit close to the current frontier models. Their performance gave Moonshot confidence that an open‑source trajectory can remain competitive against proprietary giants.

Read Also:TV Time Co-founder Launches 'Bingers' to Revive Social TV Tracking

What Kimi K3 Brings

Kimi K3 is designed not just to increase scale but to match or exceed the performance of Anthropic’s Opus 4.8. The Financial Times reports that the model will be released “in the coming days,” offering developers and enterprises the chance to fine‑tune it for specific workloads. By positioning an open‑weight system against closed‑source titans such as OpenAI’s GPT‑4, Moonshot hopes to accelerate a more transparent AI ecosystem.

Funding Surge and Valuation

In parallel with the product launch, Moonshot AI is reportedly raising a fresh capital round that would value the company at $31.5 billion. This follows a May funding event where the firm secured $2 billion at a $20 billion valuation, underscoring investor confidence in its open‑source strategy.

Open‑Source vs. Closed‑Source Debate

The announcement arrives amid a renewed industry debate over the cost and data‑privacy implications of relying on closed‑source AI services from firms like OpenAI and Anthropic. Executives across the sector are urging businesses to consider cheaper, customizable alternatives—such as models from DeepSeek, Z.ai, or Moonshot itself—especially as regulatory scrutiny over data usage intensifies. Should Kimi K3 demonstrate parity with Opus 4.8, it could trigger a shift in pricing dynamics and competitive positioning across the global AI landscape.