• AiNews.com
  • Posts
  • OpenAI to Launch Project Strawberry in ChatGPT: AI Reasoning Coming Soon

OpenAI to Launch Project Strawberry in ChatGPT: AI Reasoning Coming Soon

Futuristic digital interface representing ChatGPT with a model selection menu. The highlighted option is ‘Project Strawberry,’ featuring subtle neural network patterns in the background. A playful addition of a strawberry is subtly integrated into the design, blending with the advanced AI theme. The background features deep blues, purples, and glowing accents, evoking a high-tech, cutting-edge atmosphere while giving a nod to the strawberry concept

Image Source: ChatGPT-4o

OpenAI to Launch Project Strawberry in ChatGPT: AI Reasoning Coming Soon

OpenAI is gearing up to release its highly anticipated Project Strawberry AI reasoning model later this month, which will be integrated into ChatGPT. This new feature is expected to first be accessible to subscribers of the $20-per-month ChatGPT Plus plan.

What We Know About the Rollout

According to a report from The Information, Project Strawberry will be available as a separate option within ChatGPT, likely via the model selection menu. This early release comes ahead of previous projections, offering Plus users first access.

Project Strawberry: A Leap in AI Reasoning

Formerly referred to as Q*, Project Strawberry is designed to enhance ChatGPT’s ability to process information more thoughtfully before delivering a response. While this might slow down the response time, it aims to significantly improve accuracy and depth of answers.

Some users have already encountered a version of this reasoning model, where two responses are provided, one faster and one slower, with users asked to choose the better option. This slower response process is likely connected to Project Strawberry’s reasoning mechanism.

Access and Subscription Questions Remain

It is still unclear if Project Strawberry will require a separate subscription or if it will come with restrictions, such as limited usage or availability to only a fraction of Plus subscribers.

Fewer Errors and Improved Problem-Solving

Strawberry is poised to be a breakthrough in reasoning capabilities, allowing ChatGPT to potentially search online for more complex answers, taking extra time to respond but delivering more reliable and thoughtful outputs with fewer hallucinations.

With improved reasoning, Strawberry should excel in handling complex tasks like coding and math, offering solutions that may require multiple steps to solve. This signals a step toward more autonomous AI agents capable of performing advanced tasks.

Initial Version: Text-Only Model

The first iteration of Strawberry is expected to focus solely on text, without multimodal features like speech or video understanding. However, for text-based tasks, it is anticipated to outperform existing models with more accurate responses.

Public Availability on the Horizon

Project Strawberry is expected to appear in ChatGPT by the end of the month, quashing earlier rumors that it would remain a tool exclusively for developers to train and fine-tune other models.

Due to the high computational cost of running this advanced reasoning model, it is unlikely that free users will have access to Strawberry. Even paid users may experience rate limits as OpenAI navigates the challenges of rolling out this resource-heavy tool.

Known Challenges from Testing

In its early tests, Project Strawberry occasionally struggled with memory retention from prior conversations and exhibited slower response times, even for simple questions. Despite these hurdles, it is expected to offer a more nuanced, thoughtful interaction.

Model Updates Indicate a New Era for ChatGPT

Recently, OpenAI rebranded GPT-4 as a “legacy model,” signaling that Project Strawberry—or another new model—is likely to become the leading option for complex queries. However, it appears that we won’t be seeing GPT-4.5 or GPT-5 before major events like the upcoming election.