OpenAI up to dateUpdated version released GPT-4Turbo became available to developers and paid ChatGPT subscribers last week. When the model was launched, OpenAI said the new GPT-4Turbo offered several improvements over its predecessor, and users have found that to be true.
Starting last Thursday, an updated version of GPT-4Turbo, gpt-4-turbo-2024-04-09, reclaimed the top spot on the Large Model Systems Organization (LMSYS) Chatbot Arena, a crowdsourced open platform where users can evaluate large language models (LLMs).
In Chatbot Arena, users can communicate with two LLMs Chat side by side and compare their responses without knowing the identity of each model. These results were used to rank 82 LLMs in Chatbot Arena, including allMost PopularLLMs such as Gemini Pro, Claude3 series LLMs and Mistral-Large-2402.
As of April 13up to date Chatbot Arena update, the updated version of GPT-4Turbo is in the lead in the overall, encoding, and English categories. This means that less than a month later, Anthropic's Claude3Opus was pushed to second place in the overall category, followed by GPT-4-1106-preview, an older version of GPT-4Turbo.
These results may be attributed to gpt-4-turbo-2024-04-09's improvements in coding, math, logical reasoning, and writing skills, which outperformed a series of benchmarks used to test the proficiency of AI models. Want to compare gpt-4-turbo-2024-04-09's performance with other LLMs for yourself? You can visit the Chatbot Arena website, click the "Arena (side-by-side)" option, and select the model you want to compare.
It’s worth noting that since you know the identity of the model in side mode, you won’t be able to vote. If you want to vote and have it count towards the leaderboard, you can use the “Arena (battle)” option to compare random models. If you want to skip the test and go directly to gpt-4-turbo-2024-04-09 in ChatGPT, you need to be a ChatGPT Plus subscriber, which costs $20 per month.