Magnum 72B

Magnum 72B, from alpindale, is a new model based on Qwen2 72B, specifically trained with 55 million tokens of curated roleplay data to achieve high prose quality. In terms of performance, Magnum 72B tends to have longer response times, ranking in the 19th percentile for speed across benchmarks. Its pricing is positioned at premium levels, placing it in the 17th percentile for cost. A significant strength is its reliability, demonstrating strong performance with few technical issues, ranking in the 92nd percentile. Across benchmark categories, Magnum 72B achieved 96.0% accuracy in Email Classification, though at a higher cost and duration. Its Reasoning capabilities were moderate at 70.0% accuracy. A standout performance was observed in the Ethics benchmark, where it achieved a perfect 100.0% accuracy, making it the most accurate model at its price point and among models of similar speed. However, its General Knowledge accuracy was relatively low at 63.5%. Overall, Magnum 72B excels in ethical reasoning and classification tasks, but its speed and general knowledge breadth are areas for improvement.