Anthropic's Claude 3 Opus Tops Chatbot Leaderboard, Edging Out OpenAI's GPT-4
-
Claude 3 Opus, Anthropic's new AI model, has taken the #1 spot on the Chatbot Arena leaderboard, beating out OpenAI's GPT-4 for the first time.
-
Claude 3 Opus scored 1253 in the Elo rating system, narrowly beating out GPT-4's 1251. The scores are very close.
-
All 3 Claude 3 models (Opus, Sonnet, Haiku) are in the top 10, with Haiku reaching "GPT-4 level" according to user preferences despite being a smaller "local size" model.
-
19 of the top 20 models on the leaderboard are proprietary, suggesting open source AI still has work to do to compete with big players like Anthropic and OpenAI.
-
Meta is expected to release the open source Llama 3 model soon, which may break into the top 10 given Meta's extensive compute resources.