Key facts
- Moonshot AI's upcoming Kimi 3 model is expected to perform comparably to or better than Anthropic's Opus 4.8.
- Kimi 3 will be the largest open-weight AI model from China, with a parameter count between 2 trillion and 3 trillion.
- The model is anticipated to be released in the coming days.
- Moonshot AI is reportedly in the process of raising new capital at a valuation of $31.5 billion.
- The company had previously secured $2 billion in funding in May at a $20 billion valuation.
Moonshot AI's forthcoming Kimi 3 large language model is anticipated to rival or surpass Anthropic's Opus 4.8, according to a Financial Times report citing anonymous sources. This development positions Kimi 3 as potentially the largest open-weight AI model from China, boasting a parameter count between 2 and 3 trillion, with an expected release in the immediate future.
Moonshot AI's previous Kimi K2 models have garnered positive reception in the open-source AI market, achieving high benchmark rankings and demonstrating capabilities close to leading frontier models. The Kimi 3 release aims to further close this gap with closed-source models from major players like OpenAI and Anthropic.
This news emerges amidst a growing debate regarding the value proposition of expensive, closed-source AI models. Industry leaders express concerns about data privacy, specifically the potential for AI labs to utilize client-submitted data from products like ChatGPT and Claude. This has led to increased advocacy for alternative solutions, including companies developing their own products or customizing cheaper open-source models from entities such as DeepSeek, Z.ai, and Moonshot AI. The momentum behind open-source models is particularly fueled by their increasing ability to compete with more costly, advanced counterparts.