Ant Group Launches Ling-2.6-Flash Model, Emphasizing Token Efficiency

Ant Group launched Ling-2.6-Flash, a token-efficient large language model with 104B total parameters and 7.4B active parameters.

The model is positioned for low-cost, high-scale deployment and is now available through API access.

Ant is offering a one-week free trial and competitive pricing to attract developers and heavy users.

An earlier anonymous test version on OpenRouter saw strong usage and trended quickly, signaling market interest.