Alibaba Qianwen releases the Qwen3.8-Flash model
2026-08-26 20:38:54
According to CoinMeta, Alibaba Qianwen released the Qwen3.8-Flash model on August 26th. This is a multi-modal MOE model and also an early preview version of the Qwen4 architecture. The production version of Qwen3.8-Flash will be available through Qwen Cloud API, with a price of $0.16 per million inputs of token and $0.47 per million outputs of token. The model boasts 125 billion parameters and 51 billion n-gram embedded parameters, but only 6 billion parameters of each token are activated, achieving extremely high cost-effectiveness.
Source:Internet
This content is for market information only and does not constitute investment advice.
Follow CoinMeta official accounts to stay updated

Hot Articles
Refresh

'No longer a distant place': F2Pool Co-founder Chun Wang joins SpaceX's 2-year mission to Mars
05-22 18:25

Polymarket Targets Japan Approval Despite Gambling Laws
05-22 18:00

ZachXBT flags suspected exploit involving Polymarket's UMA adapter contract on Polygon
05-22 17:57

ZachXBT flags $520K Polymarket exploit on Polygon, team says funds are safe
05-22 17:24

Verus bridge exploiter returns 4,052 ETH, retains $2.8 million bounty: onchain analyst
05-22 17:24



