1T / 63B (MoE)
MIT
Flagship instant model in Ling family with hybrid linear attention architecture; transforms GQA layers into Lightning Linear Attention for enhanced throughput + converts remaining GQA to MLA with QK Norm and Partial RoPE for KV cache compression; 256K base context extending to 1M via YaRN; pre-trained on 29T tokens (up from 20T in Ling 2.0); reasoning approaches frontier thinking models at ~4x fewer output tokens; trained with Agentic RL in large-scale high-fidelity interactive environments; compatible with Claude Code, OpenCode, OpenClaw; leading open-source on BFCL-V4 (general tool-calling); competitive with DeepSeek V3.2, Kimi K2.5, GPT 5.2 in complex reasoning and instruction-following; high-throughput decoding for ultra-long context