Cohere: Command A

Cohere's Command A, an open-weights 111B parameter model, demonstrates strong performance across several key metrics, particularly in speed and cost-efficiency. It consistently ranks among the fastest models and offers highly competitive pricing, making it an attractive option for cost-conscious deployments. The model also exhibits exceptional reliability with a 98% success rate across benchmarks, indicating consistent and usable responses. In terms of specific benchmarks, Command A shows notable strengths in Email Classification (99.0% accuracy, 76th percentile) and Coding (88.0% accuracy, 66th percentile), aligning with its stated focus on agentic and coding use cases. Its ability to appropriately acknowledge uncertainty in the Hallucinations benchmark is also strong at 98.0% accuracy. However, a significant weakness is observed in Instruction Following, where it achieved 0.0% accuracy, suggesting a critical area for improvement. Performance in General Knowledge, Ethics, Mathematics, and Reasoning falls within the average range, indicating room for enhancement to compete with top-tier models in these domains.