Cohere: Command R+

Cohere's Command R+ model, a 104B-parameter LLM released on April 3, 2024, demonstrates a strong overall performance profile, particularly excelling in reliability. It consistently provides usable responses, achieving an exceptional 100% success rate across all benchmarks, indicating minimal technical failures. The model performs among the fastest models, ranking in the 68th percentile for speed, and offers moderate pricing, placing in the 21st percentile. Command R+ shows particular strength in practical applications like Email Classification, achieving 99.0% accuracy, placing it in the 75th percentile. It also performs well in acknowledging uncertainty, with a 94.0% accuracy in the Hallucinations benchmark. However, the model exhibits notable weaknesses in complex Reasoning (36.0% accuracy, 23rd percentile) and Instruction Following (47.0% accuracy, 44th percentile), suggesting areas for improvement in handling highly intricate tasks. Its performance in General Knowledge, Ethics, and Coding is average, hovering around the 30th-40th percentiles. The model's multilingual support for ten key languages further enhances its utility for global business operations.