DIGIVESTASI - The chatbot developed by Elon Musk's company managed to beat ChatGPT. In several tests, Grok was able to achieve a higher score than the chatbot developed by OpenAI.  Grok, which was released some time ago, announced that it had received an upgrade to version 1.5. Chatbot developer xAI explained that its latest model can understand longer documents, more complex commands and more advanced reasoning.

Entrepreneur showed that Grok 1.5 achieved 74.1% on the HumanEval benchmark. It consists of 164 programming problems and is not included in the AI model training.  GPT-4 achieved 67%. Gemini Pro, a Google chatbot, had a score of 71.9%. 

On the MMLU test, Grok obtained almost the same result (83.7%) as Gemini in knowledge of 57 basic and advanced subjects. The latest model scored 81.3% recorded on Monday (April 1, 2024). Mr. Grok scored 50.6% on a math test that included competitive questions for grades K-12. The score in this category is still below the GPT-4 or 52.9%.

It's no secret that Musk has big ambitions for chatbots. The X report states that the new Grok will outperform existing AI on every metric. He also said that updates to Grok will be ongoing. "Grok 2 is in training," he said. Grok was announced late last year and is gearing up to be available to X users. However, the chatbots can only be used with paid services within the app. Grok costs $16 per month and includes X-Premium features.

Find more news and articles on Google News