While the arrival of GPT-4 made ChatGPT significantly smarter and more accurate in many regards, there’s one bit that hasn’t changed a lot, and it’s the way it communicates. It tends to be rather wordy and verbose, even in situations where it’s not really required. Now, OpenAI is rolling out a new version of GPT-4 Turbo that’s apparently “more direct” and conversational.
OpenAI has rolled out a significant upgrade to ChatGPT. Premium users of the chatbot that are subscribed to ChatGPT Plus, Team, or Enterprise plans now have access to an enhanced version of the GPT-4 Turbo model, labeled “gpt-4-turbo-2024-04-09.” This updated model promises improvements in various areas such as writing, math, logical reasoning, and coding, along with a more current knowledge base. Unlike its predecessor, which had a cut-off in April 2023, this version was trained on publicly available data up to December 2023. The free version of ChatGPT is trained on data up to January 2022, so April 2023 was already pretty good, but the fact that it was bumped up to December means that ChatGPT premium users are using very, very recent data—the data is only four months old, after all.
OpenAI says that interactions with ChatGPT utilizing the new GPT-4 Turbo will result in more concise, conversational responses. What this means is that you can expect less wordy answers to simple queries, as well as more to-the-point answers. We’ll have to see how the updated model compares to previous ones in this regard, but it certainly looks promising. The news comes after OpenAI started allowing people to use ChatGPT without creating an account first, though that mode is limited to the same GPT-3.5 model used for free acccounts.
If you want to check it out, just fire up ChatGPT if you’re a paid user and shoot some prompts at it. If you’re a free user, you’ll have to pay for Plus, as the older model is still based on the old reliable GPT-3.5.
Source: TechCrunch
