Open AI is working on “Strawberry” New AI Models

Chat GPT developer Open AI is developing a new kind of reasoning AI models with the project name “Strawberry” that can be used for research, according to a report by Reuters. Strawberry was apparently earlier known by the name “Q” and would be considered a breakthrough within Open AI.

The plan is for the new Strawberry models to not only be able to generate answers based on instructions, but also to be able to plan ahead by navigating the internet independently and reliably perform what Open AI calls “deep research.”

How Strawberry works under the hood remains unclear; it is also unknown how long the technology may be from completion. In a comment to Reuters, an OpenAI spokesperson says continued research into new AI opportunities is ongoing within the industry. However, the spokesperson did not say anything specific about Strawberry in particular.

What is OpenAI Strawberry model?

Strawberry, reportedly an upgrade of OpenAI’s earlier Q* project, is believed to be a significant leap in AI development. Q* had stirred internal debates and was linked to the temporary dismissal of CEO Sam Altman in November 2023, according to insiders.

Bloomberg reports that OpenAI leaders have outlined a progression framework for AI capabilities, currently placing the company at the initial “Reasoners” level with ChatGPT-4. The stage is characterized by AIs that can solve basic problems as effectively as a highly educated human without tools. The subsequent stages include “Agents,” where AI can autonomously perform tasks over several days, and “Innovators,” capable of devising new technologies.

The ultimate tier, “Organizations,” represents AI systems that can work with comprehensive autonomy and complexity. These stages sketch the path towards achieving full AGI, demonstrating OpenAI’s vision for the future of AI technologies.