Ant Baoling Open-sources Trillion-Parameter Comprehensive Flagship Model Ling-2.6-1T
Baoling Large Model officially open-sources Ling-2.6-1T, a trillion-parameter comprehensive flagship model designed for complex tasks. Relying on a hybrid architecture of MLA and Linear Attention, the model achieves a high intelligence-efficiency ratio with extremely low Token consumption, placing its comprehensive intelligence on par with GPT-5.4 (Non-Reasoning). It reaches open-source SOTA on multiple execution-based benchmarks including AIME26, SWE-bench Verified, and BFCL-V4, possessing strong Agent execution, code generation, and long context understanding capabilities.
Baoling Large Model officially open-sources Ling-2.6-1T, a trillion-parameter comprehensive flagship model designed for complex tasks. Relying on a hybrid architecture of MLA and Linear Attention, the model achieves a high intelligence-efficiency ratio with extremely low Token consumption, placing its comprehensive intelligence on par with GPT-5.4 (Non-Reasoning). It reaches open-source SOTA on multiple execution-based benchmarks including AIME26, SWE-bench Verified, and BFCL-V4, possessing strong Agent execution, code generation, and long context understanding capabilities.
This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.