OpenAI Resolves "Laziness" Issue In GPT-4 Turbo Model And Introduces Lower Pricing

File Photo

New Delhi: In a move to enhance user experience and to operate hussle-free, OpenAI has introduced a significant update to its AI models and pricing. The updates bring several new features including lower pricing for the popular GPT-3.5 Turbo model, improved performance for GPT-4 Turbo, and upgraded text embedding models.

The per token pricing of GPT-3.5 has been slashed by OpenAI by 50% for input and 25% for output. This model is crucial for applications such as ChatGPT and is widely recognized in the conversational AI industry. The price reduction aims to make the API more affordable for developers working on text-intensive products that involve tasks like document or book analysis. As open-source models advance, the lowered pricing strategy is designed to retain customers. ()

GPT-3.5 Turbo has been updated to version 0125 featuring unspecified "improvements." Given that the previous version was 0613. Some individuals might have anticipated more detailed information on the advancements made by OpenAI. 

A new preview model version 0125 has been introduced for GPT-4 Turbo. OpenAI states that this version improves the completion of tasks such as code generation addressing the issue of "laziness" where the model might avoid completing its work thoroughly.

The objective is to enhance performance on technical tasks compared to earlier versions. While GPT-4 with vision capabilities is currently in preview, it is expected to have a broader launch in the upcoming months.

From a technical standpoint the improved text embedding models offer more effective semantic representations of language for researchers and engineers. Although not groundbreaking, OpenAI's recent adjustments showcase the company's gradual and continuous efforts to enhance its AI offerings.