The great AI repricing
While SpaceX failed to launch its latest rocket after multiple engine failures late last week, China's Moonshot AI pulled off a moonshot of its own. The startup unveiled its latest AI model, Kimi K3, on Friday and shook US markets for two reasons.
First, Kimi K3 matched or outperformed Anthropic's Fable 5 and OpenAI's GPT-5.6 Sol on front-end coding benchmarks compiled by AI evaluator Arena. That is the performance leg of the story.
Second, it outperformed Anthropic's Opus 4.8 — the company's flagship model before Fable 5 — while costing around 40% less.
Read the full article here.
Author

Ipek Ozkardeskaya
ipekScope
Ipek Ozkardeskaya began her financial career in 2010 in the structured products desk of the Swiss Banque Cantonale Vaudoise. She worked in HSBC Private Bank in Geneva in relation to high and ultra-high-net-worth clients.


















