Tech

Moonshot AI’s Kimi K3 set to challenge Anthropic as valuation targets $31.5 billion

Anonymous sources tell the Financial Times the new 2 trillion to 3 trillion parameter model will rival Opus 4.8, coinciding with a funding round that values the company at $31.5 billion.

Author
Owen Mercer
Markets and Finance Editor
Published
Draft
Source: TechCrunch · original
Moonshot’s upcoming Kimi 3 is expected to close the gap with Anthropic’s Opus 4.8
Chinese lab prepares largest open-weight model release amid fresh capital raise

Chinese artificial intelligence lab Moonshot AI is preparing to launch Kimi K3, its largest open-weight model to date, featuring a parameter count between 2 trillion and 3 trillion. According to a report by the Financial Times citing anonymous sources, the model is expected to perform at par with or surpass Anthropic’s Opus 4.8. The release is anticipated in the coming days, marking a significant step in the company’s efforts to compete with Western frontier models.

The announcement coincides with Moonshot AI’s efforts to secure fresh capital in a funding round that would value the company at $31.5 billion. This represents a substantial increase from May, when the lab raised $2 billion at a $20 billion valuation. The timing of the model release and the capital raise suggests a strategic push to capitalise on growing demand for high-performance open-source alternatives in the global AI market.

Kimi K3 aims to further close the performance gap between open-weight and closed-source models. Previous iterations, specifically the Kimi K2 models, have ranked highly on industry benchmarks and demonstrated capabilities that are increasingly comparable to the latest frontier offerings. The new model is positioned to address industry concerns regarding the high costs and data privacy risks associated with proprietary systems from providers such as OpenAI and Anthropic.

Executives are increasingly recommending cheaper open-source alternatives, including models from DeepSeek, Z.ai, and Moonshot, for corporate training purposes. This shift is driven by fears that closed-source labs may extract client-submitted data to improve their proprietary products. As Chinese models continue to narrow the performance divide with expensive, closed-source counterparts, the market sentiment is shifting towards more transparent and cost-effective solutions.

While the Financial Times reports that Kimi K3 will rival Opus 4.8, these performance claims are based on anonymous sourcing and have not been independently verified. The exact release date remains fluid, described only as occurring in the coming days. The final $31.5 billion valuation is also contingent on the successful completion of the current funding round.

Continue reading

More from Tech

Read next: Kimi AI launches K3 model with focus on agentic coding and parallel task execution
Read next: OpenAI’s hardware pivot risks deepening losses as smart speaker market shrinks
Read next: Suno integrates generative AI music into Apple’s iMessage