OpenAI has been creating some really powerful AI models for its tools, and they’ve been able to keep companies like Google on their toes. Not too long after releasing GPT-4o mini, we got word about OpenAI’s next model codenamed Strawberry. From the sounds of it, it seems that the model will be a smarter model, and it will only cost you time (and maybe some money).
If you don’t know what GPT-4o mini is, it’s a stripped-down version of the GPT-4o. While it’s not as powerful, it’s still pretty capable, and it’s more powerful than GPT-3.5 Turbo. It’s not the model that’s powering ChatGPT for free users. GPT-3.5 Turbo is still available for people who use the ChatGPT API.
OpenAI’s new model is codenamed Strawberry
Right now, we don’t know too much about this new mode, but we’re going to find out more soon. According to the report, the company is going to launch this model in two weeks. This goes with the company’s strategy of launching often.
As for this model’s capabilities, while companies tout the speed of their models, Strawberry won’t exactly be a speed demon. In fact, we expect it to be slower than most models on the market, and that’s a good thing. Strawberry could be a more reasoning-focused model. As such, the model might actually take longer to deliver answers than other models. As per the report, it could take between 10 and 20 seconds to actually start giving you your answers.
The model will take extra time before giving you your answer. This might reduce the number of inaccurate facts it spits out. It makes sense. Imagine a person taking 10 seconds to think and double-check and second guess themselves before giving you an answer.
The model is expected to remember your past conversations and give you more personalized responses.
As for the cost, that’s still up in the air. Usage of this model could be limited to a certain amount of uses per hour. Just like with other models, you might be able to switch between Strawberry and other models.
2024-09-12 15:11:02