Arcee AI: Trinity Mini

Arcee AI's Trinity Mini, a 26B-parameter sparse mixture-of-experts model, demonstrates exceptional speed and competitive pricing, consistently ranking in the Infinityth percentile across eight benchmarks for both metrics. Its reliability is strong, with a 96% success rate across benchmarks, indicating minimal technical failures. The model excels in classification tasks, achieving 98.0% accuracy in Email Classification, and shows robust performance in Reasoning (90.0% accuracy) and Coding (87.0% accuracy). However, a significant weakness is its complete failure to acknowledge uncertainty, scoring 0.0% accuracy in Hallucinations, indicating it does not appropriately select "I don't know" for fictional concepts. Similarly, its Instruction Following capability is non-existent at 0.0% accuracy, suggesting it struggles with complex multi-step directives. While General Knowledge and Mathematics scores are respectable, its Ethics performance is notably low at 82.0% accuracy (18th percentile), coupled with a very long duration.