Moonshot’s Kimi 3 Aims to Compete with Anthropic’s Opus 4.8

According to a Financial Times report citing anonymous sources, the latest iteration of the Kimi model series from the Chinese AI lab Moonshot AI is expected to rival or surpass the performance of Anthropic’s Opus 4.8.

Moonshot’s Kimi K2 models have garnered positive recognition in the open-source AI community, achieving impressive results on numerous benchmarks and demonstrating capabilities that compete strongly with the most advanced models available.

The anticipated Kimi K3 model is said to push advancements even further, aiming to close the performance gap with proprietary models from firms like OpenAI and Anthropic. As reported by the FT, the Kimi K3 is set to be the largest open-weight AI model developed in China, expected to feature between 2 trillion and 3 trillion parameters, with a launch anticipated “in the coming days.”

In addition, Moonshot is reportedly in the midst of raising new funds, aiming to reach a valuation of $31.5 billion. The company had previously raised $2 billion in May at a valuation of $20 billion.

This news comes during ongoing conversations about the rationale behind investing in AI labs such as OpenAI and Anthropic, particularly given the high costs associated with their proprietary models. Industry leaders are voicing concerns that these AI labs might exploit data provided by clients for their AI services, including ChatGPT and Claude.

Executives are advocating for their alternative offerings, urging businesses to explore more budget-friendly open-source models, such as those developed by DeepSeek, Z.ai, or Moonshot, adapting them to fit their specific requirements. This viewpoint has gained momentum, especially as open models from China continue to compete effectively with their more expensive equivalents.