The company’s engineers tested two AI models from Openai: O1, GPT-4O and Claude 3.5 Sonnet from Anthropic. The models were evaluated using a SWE-Lancer comparison of 1,400 tasks for programmers with the Upwork Freelance region.
During the test, models are prohibited from internet access that replicated network solutions and excluding the possibility of “deception .. The researchers found that although AI models show “some competence ,, they could not replace even new programmers.
Models made mistakes and “had almost ever understood the context, which led to false or inadequate solutions. At the same time, the Claude 3.5 Sonnet model showed the best results, but most of the answers and this nervous network was still wrong.
Researchers will now appear until the end of 2025, contradicting the expression of the models that can solve programming problems, which are fundamentally contradictory by the expression of the General Manager of Openai Sam Altman, that AI cannot even write a simple code at the moment and even more replacement experts.
Source: Ferra
I am a professional journalist and content creator with extensive experience writing for news websites. I currently work as an author at Gadget Onus, where I specialize in covering hot news topics. My written pieces have been published on some of the biggest media outlets around the world, including The Guardian and BBC News.
