Mark RadarMARK RADAR
中 EN
Event File AI

Moonshot AI’s Yang Zhilin Launches Frontier Kimi K3 Model

1 reports · First detected 2026-07-24 · Last active 2026-07-24

Yang Zhilin, a Tsinghua University alumnus with a doctorate from Carnegie Mellon University, co-founded Beijing-based Moonshot AI in 2023. The Alibaba-backed startup gained prominence with its long-context Kimi assistant. Its advances are closely watched as a measure of China’s ability to narrow the generative AI gap despite US restrictions on exports of advanced chips used to train frontier models.

Moonshot AI launched Kimi K3 on July 16, 2026, describing the 2.8-trillion-parameter system as the world’s largest open-weight model. It supports a 1-million-token context window and delivers performance approaching Anthropic’s Fable, the company said. API access costs $3 per million input tokens and $15 per million output tokens, while Moonshot plans to release the full model weights by July 27.

All Coverage

1 original reports
FT.COM 2026-07-24
Yang Zhilin, the rock star founder behind China’s Moonshot AI

The Backstory

The history behind this event
Moonshot AI Founder’s China Return Fuels Talent Debate as Kimi K3 Emerges2026-07-22 · 1 reports · similarity 0.81

Moonshot AI founder Yang Zhilin’s decision to leave Silicon Valley and build an artificial-intelligence company in China has become a closely watched example of global technology talent shifting between the world’s two largest economies. Moonshot’s rise has also sharpened debate over whether US immigration policy can retain international researchers as China expands its domestic AI ecosystem and challenges American leadership in advanced models.

Yang has since led Moonshot AI in launching Kimi K3, a large language model that brought the company broader international attention and made his return central to discussion about China’s ability to attract experienced founders. The available report did not specify Kimi K3’s launch date, development cost or parameter count. Attention is now extending beyond the model’s performance to talent retention, founder mobility and the longer-term impact of US and Chinese technology policies.

Mark Radar｜MARK RADAR
Past Issues・by MarkReadFintech
All times are in Taipei time (GMT+8)