The company aims to leverage Strawberry to assist software and AI developers, enable predictive planning for autonomous internet searches, and support in-depth research, along with other potential applications. (Source: Image by RR)

‘Strawberry’ Expected to Offer Faster Reasoning for Design, Math and Programming

OpenAI plans to release its new AI model, codenamed “Strawberry,” within the next two weeks, ahead of the previously reported fall release. Designed to improve logical reasoning, Strawberry will be integrated into ChatGPT, enhancing its ability to solve complex problems in areas like design, mathematics, and programming. As noted in the-decoder.com, a unique feature of the model is its 10 to 20-second “thinking” period before responding, allowing it to tackle tasks it hasn’t encountered before.

Strawberry uses a specialized form of “post-training” to adapt pre-trained models for specific tasks, which could reduce the need for additional prompts. During internal testing, the model reportedly scored over 90% on mathematical mastery tasks. It is also being used to provide training examples for OpenAI’s next-generation AI model, codenamed “Orion.” Although part of ChatGPT, Strawberry will initially only process text, unlike other models that handle multiple input types.

The model is intended to support software and AI developers, enable autonomous internet searches, and assist in in-depth research. However, early testers have reported that Strawberry can sometimes take too long to process simple queries, and it hasn’t yet fully adapted to user preferences like specific coding formats. Some testers feel that the model’s slight improvements over GPT-4 don’t justify the extra 10 to 20-second wait time.

Pricing for OpenAI’s upcoming “Strawberry” model is expected to differ from that of ChatGPT, potentially introducing new rate limits and subscription tiers to cater to various user needs. Speculation suggests that OpenAI may offer a premium pricing model for faster response times, appealing to users who require high-speed processing for complex tasks. Additionally, OpenAI is reportedly considering raising the prices of its advanced AI models overall, which could reflect the increasing sophistication and capabilities of the technology. This strategy aligns with the company’s need to monetize its evolving models, particularly as demand grows for AI systems that can handle more specialized and advanced tasks.

read more at the-decoder.com