Ant Group's Ling 3.1 Flash Ranks Fifth on Intelligence Index With 560B Mixture-of-Experts Architecture

Ant Group's InclusionAI lab has released Ling 3.1 Flash, a 560-billion-parameter reasoning model that achieved a score of 41 on the Artificial Analysis Intelligence Index, jumping from its predecessor's score of 20 and placing it among the top five models in the benchmark. The model activates approximately 25 billion parameters per token, operates with a 1-million-token context window, and is priced at $0.30 per million input tokens and $0.90 per million output tokens through the InclusionAI API.