OpenAI’s highly anticipated AI model, Strawberry (officially o1), has launched, marking a significant milestone in AI development.
This model is designed to handle complex tasks in areas like coding, mathematics, and scientific research—bringing OpenAI one step closer to achieving artificial general intelligence (AGI).
Strawberry’s capabilities have the potential to reshape industries that rely on:
- Advanced Reasoning
- Autonomous Research
- Industry Impact
However, the road to launch wasn’t without challenges. OpenAI faced months of internal conflict, including the temporary ouster of CEO Sam Altman over concerns surrounding AI risks.
Despite these setbacks, the company has moved forward, and Strawberry is now positioned to make a lasting impact on AI’s role across various sectors.
How It All Began: The Cryptic Tweet
The first hints of this new AI model came in mid-2024 when OpenAI CEO Sam Altman posted a cryptic tweet featuring a photo of strawberries with the caption, “I love summer in the garden.”
i love summer in the garden pic.twitter.com/Ter5Z5nFMc
— Sam Altman (@sama) August 7, 2024
This seemingly simple post quickly sparked interest in the AI community, as it was seen as a hint at a major upcoming project. Internally, the project was known as Strawberry, though specifics were kept under wraps.
The mystery surrounding the project intensified in November 2023 when it was revealed that Strawberry was linked to a secret AI discovery. OpenAI researchers raised concerns about the project’s potential risks, leading to Altman’s temporary removal as CEO, Reuters reported.
The letter highlighted the risks associated with AI’s rapid development and was one of the catalysts behind Altman’s short ouster as CEO. Reuters further reported that this move set off a wave of protests within the company, as over 700 employees threatened to resign in support of Altman.
Shortly after, Altman was reinstated, and the company’s focus shifted back to preparing for the launch.
From Hints to Headlines: The Journey to Strawberry
Strawberry is officially launched as part of OpenAI’s efforts to push the boundaries of AI capabilities.
Unlike traditional AI models, Strawberry is designed to go beyond basic conversation, offering advanced reasoning and the ability to handle complex, multi-step tasks.
See more: OpenAI Supports California Bill For Labeling AI-Generated Content
The model is expected to perform independent research and plan tasks, bringing OpenAI closer to achieving AGI.
According to earlier reports from Reuters, Strawberry can handle multi-step functions from a single query and “plan sufficiently to autonomously navigate the internet and reliably perform what OpenAI describes as ‘deep research.'”
These advancements brought the model closer to OpenAI’s goal of developing fully autonomous AI agents and are offering the following:
Key Features and Capabilities of Strawberry
- Advanced Reasoning: Strawberry enhances AI’s ability to think through problems and provide reasoning-focused responses, a capability that sets it apart from earlier models.
- Autonomous Research: The model can navigate the internet, conduct deep research, and plan ahead, showcasing abilities that resemble human decision-making processes.
- Industry Impact: With its advanced capabilities, Strawberry is poised to impact industries such as software engineering and scientific research by helping solve complex, multi-layered problems.
OpenAI’s chief, Sam Altman, went to X to celebrate this monumental achievement. “Extremely proud of the team; this was a monumental effort across the entire company. Hope you enjoy it!”, he posted.
extremely proud of the team; this was a monumental effort across the entire company.
hope you enjoy it!
— Sam Altman (@sama) September 12, 2024
Despite its potential, Strawberry does face some limitations. The model, while powerful in reasoning, is slower than previous versions, taking 10-20 seconds to process complex queries.
This slower response time may affect its broader adoption, especially in business environments where speed is crucial. Nevertheless, OpenAI is positioning Strawberry as a model for mission-critical tasks where accuracy and reasoning are more important than speed.
What’s in the OpenAI’s Model?
OpenAI’s new AI model, Strawberry (o1-preview), has launched, bringing significant advancements in reasoning and problem-solving.
1. Impressive Performance
Strawberry excels in various domains, ranking in the 89th percentile in competitive programming and scoring 83% on an International Mathematics Olympiad qualifying exam. This marks a substantial improvement over the GPT-4o model, which scored only 13%.
2. Enhanced Safety Features
Strawberry includes improved safety measures, scoring 84 out of 100 on jailbreaking prevention, compared to GPT-4o’s 22. OpenAI is collaborating with safety institutes to further enhance the model’s security.
3. A o1-Mini
Alongside Strawberry, OpenAI has released o1-mini, a cost-effective version for developers needing advanced coding capabilities. Priced 80% lower than Strawberry, o1-mini is available to ChatGPT Plus and Team users with a 50-message daily limit. API users can begin testing now, with broader access planned for the future.
What’s Next for AI?
OpenAI’s Strawberry model is a significant milestone in the company’s efforts to develop AI that can think and analyze like humans.
While it still has some limitations, Strawberry represents a crucial step toward creating autonomous AI agents capable of tackling complex, multi-step tasks more efficiently.
As OpenAI continues to develop and refine this technology, the world will watch closely to see how it is integrated into society for a better future.
Co-founder and Chief growth officer