Every FLOP Counts: Scaling a 300B Mixture-of-Experts LING LLM without Premium GPUs

Benchmark Model Rank Results
instruction-following-on-multichallengeLing-3.1-flash#1Accuracy: 69.78
vulnerability-reproduction-on-cybergymLing-3.1-flash#2Success Rate: 87.9
code-generation-on-terminal-bench-4-0Ling-3.1-flash (LingCode)#5Accuracy: 46.03