Alibaba's 14B Qwen 2.5 instruct model — strong coding, math, and instruction-following at a mid-size, cost-efficient deployment point.
Qwen2.5-14B-Instruct FC — estimated from Qwen2.5 family (between 7B and 32B scores)
Qwen2.5-14B-Instruct — Open LLM Leaderboard proxy
rank #173, IFEval=81.58 BBH=48.36 MATH=54.76 GPQA=9.62 MUSR=10.16 MMLU-Pro=43.38
