Mistral Large 2407

Mistral Large 2407, Mistral AI's flagship model, demonstrates a strong overall performance profile. It performs among the fastest models, ranking in the top tier for speed (74th percentile across 8 benchmarks). The model offers moderate pricing, positioned at the 31st percentile. Notably, Mistral Large 2407 exhibits exceptional reliability, achieving a 100% success rate across all 8 benchmarks, indicating minimal technical failures and consistent response delivery. Across benchmark categories, Mistral Large 2407 excels in "Hallucinations (Baseline)" with perfect 100% accuracy, demonstrating its ability to appropriately acknowledge uncertainty. It also shows strong performance in "General Knowledge" (99.5% accuracy) and "Email Classification" (99.0% accuracy). Its proficiency extends to supporting dozens of human languages and over 80 coding languages, further enhanced by a 131072 context length for precise information recall. While "Ethics" and "Mathematics" show solid performance (99.0% and 88.0% accuracy respectively), "Instruction Following" (62.0% accuracy) and "Reasoning" (58.0% accuracy) represent areas with room for improvement compared to its other strengths. "Coding" performance is respectable at 82.0% accuracy.