Perplexity CEO: May Use Kimi K2 for Post-Training in the Future
According to a report by Jinse Finance, Aravind Srinivas, CEO of the US-based AI search startup Perplexity, stated on social media that due to the strong performance of the Kimi K2 model, the company may use K2 for further training in the future. Previously, DeepSeek R1 was also used by Perplexity for model training. K2 is a trillion-parameter open-source model recently released by Moonshot AI’s Kimi, emphasizing coding capabilities and general agent task performance.
Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.
You may also like
Golden Sun Health Technology Proposes Bylaw Amendments to Reflect Class B Share Changes
BTC breaks through 83,000 USDT, with a 24-hour increase of 1.12%
Super Micro Computer Q1 preliminary earnings soar with revenue up 27% year-on-year, stock surges 11% after hours
Super Micro Computer has released its preliminary Q1 results for the 2027 fiscal year, projecting revenue between $18.5 billions and $19.5 billions, significantly exceeding the previous guidance of $14.0 billions to $15.5 billions and analyst consensus of $15.23 billions—an increase of approximately 27%. Gross margin is preliminarily expected to be between 12.0% and 12.5%, also higher than earlier guidance. Following the announcement, the company's share price surged 11% in after-hours trading. The company will hold its full quarterly earnings call on November 3, and final figures may be subject to adjustment.
Crypto prices
More
