Command R+ Reaches 6th Spot, First Open-Weights Model in AI Arena

The AI community is abuzz with the latest results from the Arena, where @cohere's Command R+ has made a significant leap to the 6th spot, tying with GPT-4-0314, as per 13K+ human votes. This achievement marks Command R+ as the best open model and the first open-weights model currently on the leaderboard. Experts highlight the importance of benchmarks like the Chatbot Arena for evaluating large language models (LLMs), where Command R+ has been recognized for its performance, even without testing additional capabilities such as RAG & tool use. Additionally, Command R+ has received commendations for its scores on MMLU and GSM8K, further establishing its position. Meanwhile, the AI field continues to evolve with the introduction of new models like MistralAI's 7x22B MoE (170B) and updates on other notable models like AnthropicAI's Claude 3 Opus and Gemma-1.1-7B, with Command R+ being highlighted as the best public (non-commercial use-case only) LLM judging by the lmsys benchmark.