Ant Group Unveils Ling-2.6-Flash: A Major Leap in AI Efficiency

AI Briefing

• – Ant Group unveils Ling-2.6-flash, a new AI model prioritizing efficiency and real-world application with 104 billion total parameters and 7.4 billion active parameters.
• – The model uses a sparse Mixture-of-Experts (MoE) architecture to deliver high intelligence at lower cost and latency compared to peers.
• – Ling-2.6-flash is designed to address limitations of existing models that rely on generating excessive tokens to achieve higher benchmarks.