results.csv

Model,Idiom_Translation,Lexical_Ambiguity,Terminology_Localization,Tense_Consistency,Zero_Pronoun_Translation,Cultural_Safety,OverAll,Type
GPT-4o,5.04,5.5,4.54,5.22,5.56,4.7,5.09,industry APIs
DeepSeek-v3,5.14,5.52,5.12,5.46,5.06,4.66,5.16,industry APIs
DeepSeek-R1-70B,3.82,4.1,4,4.9,4.6,4.74,4.36,open-source
LLaMA3-70B,2.76,2.28,3.64,4.34,3.82,4.62,3.58,open-source
Qwen3-32B,4.5,4.32,4,5.1,5,4.42,4.56,open-source
Qwen3-14B,4.24,3.9,4.12,4.9,4.68,4.42,4.38,open-source
Qwen3-8B,4.1,3.78,3.32,4.66,3.62,4.26,3.96,open-source
LLaMA3-8B,2.64,2.56,2.7,3.98,2.84,4.04,3.13,open-source
ChatGLM4-9B,3.12,3.62,3.4,4.34,3.32,4.3,3.68,open-source
Google Translation,4.72,4.52,4.54,4.24,4.82,4.52,4.56,open-source
IFLYTEK Translation,2.66,2.46,3.12,4.1,2.2,4.04,3.1, translation-specialized
Seed-X-PPO-7B,5.42,5.46,4.52,4.64,3.68,4.2,4.65, translation-specialized
Seed-X-Instruct-7B,3.8,4.3,4.16,4.58,3.24,3.92,4, translation-specialized
GemmaX2-28-9B,3.6,3.26,3.66,4.88,3.28,4.34,3.84, translation-specialized