Fierce competition continues in the world of artificial intelligence. Although OpenAI’s GPT-4 based model is considered to be the most powerful model of recent times, new tests have revealed that this is behind us. So which is the best AI model?
Claude-3 vs ChatGPT-4: The best AI model
The LMSys Chatbot Arena Leaderboard, which tracks the performance of artificial intelligence systems, revealed that the Claude-3 Opus model developed by Anthropic received the highest score. This means that GPT-4, the leader for many years, has fallen behind.
The Leaderboard was evaluated by judges during “chat battles”. The judges, who tested chatbots by entering various commands from different topics, used the Elo rating system for comparison.
Claude-3 Opus came out on top with 1253 Elo points in the latest rating update. In second and third place were the OpenAI GPT-4 preview models with 1251 and 1248 points, respectively.
In the rest of the ranking, Google’s Gemini Pro came fourth with 1203 points and Anthropic’s earlier Claude-3 Sonnet came fifth with 1198 points. Chinese company Alibaba’s Qwen1.5 came in ninth place with 1148.
Here is the list of the best AI models:
- Claude 3 Opus – 1253
- GPT-4-1106-preview – 1251
- GPT-4-0125-preview – 1248
- Bard (Gemini Pro) – 1203
- Claude 3 Sonnet – 1196
- GPT-4-0314 – 1193
- Claude 3 Haiku – 1181
- GPT-4-0613 – 1174
- Mistral-Large-2402 – 1153
- Qwen1.5-72B-Chat – 1153