OpenAI Aims To Release New AI Model, ‘Strawberry’, In Fall (2 minute read)
OpenAI is planning to release a new AI product called “Strawberry” in the fall. It will feature advanced reasoning capabilities, such as the ability to solve previously unseen math problems, and can perform high-level tasks like deve—–loping market strategies.
“New details on OpenAI’s “Strawberry” are out—and there’s a new model called Orion.
|
|
The Information dropped some fun new intel on Project Strawberry—it sounds like OpenAI is trying to launch their new reasoning chatbot as early as this fall(!!!). |
According to two insiders, Strawberry can do a bit more than we thought: |
- Solve novel math problems it’s never seen before.
- Help you with your programming, or “think deeply” to plan your product marketing strategies.
- Employees at OpenAI have even used it to solve the NYT Connections! (Where’s my “purple first” crew at?! IYKYK)
|
Don’t get too excited…it sounds like the version of Strawberry we’ll get to use will be a smaller, “distilled” version of the original. It also remains too soon to say if Strawberry-lite will launch this year, or get pushed to 2025. |
Knowing OpenAI… plan for 2025. |
Here’s how Strawberry works: According to the report, it sounds like OpenAI lets Strawberry “think” longer than we’re used to letting AIs think. |
|
This means little-ole Strawbs might take longer to get you an answer—buuut that answer might better understand what you’re really asking for. |
This news is not all about Strawberry, though—according to The Information, there’s another chatbot OpenAI is building named “Orion.” |
Orion is supposed to be OpenAI’s answer to all the new frontier chatbots trying to unseat it (think Gemini 1.5 Pro, Grok 2, Claude 3.5 Sonnet). |
OpenAI is apparently using the big, non-distilled version of Strawberry to generate synthetic training data for Orion, which some think could help reduce hallucinations.” |
|
“OpenAI reportedly preparing to launch new ‘Strawberry’ AI model
|
OpenAI is reportedly on the brink of launching its new AI model, “Strawberry,” which could advance towards the goal of artificial general intelligence. Previously known internally as Q*, Strawberry is expected to surpass its predecessors by addressing complex mathematical problems, developing sophisticated marketing strategies, and solving intricate puzzles like the New York Times’ “Connections.” It boasts a MATH benchmark score exceeding 90%, outperforming both GPT-4 and GPT-4o, and positioning OpenAI ahead of its competitors. |
|
The model’s development has caused internal disruption at OpenAI, including a brief period of CEO Sam Altman’s removal, due to concerns about Strawberry’s potential to achieve significant breakthroughs and the associated risks of AGI. |
A release is anticipated for the fall, with future scenarios potentially allowing ChatGPT users to switch between Strawberry and other models based on the urgency of their needs. Read more.” |
|
|
“Our Report: Following rumors in July that OpenAI was working on Project Strawberry—designed to enhance the reasoning capabilities of AI models, a step towards Artificial General Intelligence (AGI) which means it can perform tasks at a human level—and a cryptic post a couple of weeks ago from CEO, Sam Altman, of a pile of strawberries, insiders have reported that Project Strawberry is likely to launch in the fall, and possibly power ChatGPT-4’s successor, “Orion.” |
Key Points: |
- Previously called Q*, and the brainchild of former chief scientist Ilya Sutskever, Strawberry can (reportedly) solve unseen math problems, has advanced programming skills, and can respond to more subjective topics.
- OpenAI’s main aim for Strawberry is to enhance its self-learning and logical thinking capabilities so it can evolve and adapt without human guidance, and make reasoned decisions based on data.
- They’re using the larger version of Strawberry (which generates high-quality ‘synthetic’ training data, reducing their need to use text and images from the internet, reducing hallucinations) to train the new ChatGPT: Orion.
|
Why you should care: OpenAI has reportedly already demoed Strawberry to US national security officials, to appear more transparent with policymakers who fear the rise of AI, which many believe is an initiative to quell the rising concerns over transparency and safety after a tumultuous few months that saw multiple safety leaders quit, citing safety concerns.” |
|
 |
Image source: Midjourney |
|
“The Rundown: OpenAI researchers are preparing to launch a new AI model, code-named Strawberry (previously Q*), that demonstrates superior reasoning capabilities in solving complex problems, according to a new report via The Information. |
The details: |
- Project Strawberry could be integrated into ChatGPT as soon as this fall, marking a significant leap in AI intelligence.
- Given extra “thinking” time, Strawberry can tackle subjective topics and solve complex puzzles like the New York Times Connections.
- OpenAI is using Strawberry to generate high-quality training data for another secretive upcoming LLM, reportedly code-named Orion.
- The new AI model could enhance OpenAI’s development of AI agents, potentially automating multi-step tasks more effectively.
|
Why it matters: If Strawberry lives up to the leaks, it could mark a significant leap in AI reasoning capabilities, potentially advancing OpenAI towards Stage 2 of its five-level roadmap to AGI. With ChatGPT reported to gain these capabilities this fall, we’re likely on the verge of seeing the next major wave of AI disruption.” |
Is it Strawberry Season?
|
|
“Details about OpenAI’s “Strawberry” – a rumored model that could take AI reasoning to the next level – have finally been revealed. |
Here they are: |
- Sources told The Information OpenAI might integrate Strawberry into ChatGPT, instead of releasing it as a standalone model
- It’s reportedly so powerful that a team showed it off to American national security officials this summer
- It would be able to perform high-level math and logic problems — even those it was never trained on, although it’d take longer to generate results
- It could be released as soon as this fall, with a new model code-named Orion coming at a later date
- The company is struggling to raise more capital, so this could be just the boost it needs to power through”
|
|
WHY IS EVERYONE TALKING ABOUT STRAWBERRIES? 
|
 |
Image made with Grok-2 |
|
Q: Why did the strawberry win the lottery? |
A: It was berry lucky.  |
What’s up? OpenAI is reportedly on the brink of launching an advanced AI model, initially known as Q* but now internally dubbed “Strawberry.” |
What is Strawberry? The Strawberry model promises to outperform current AI systems by tackling problems and tasks that have previously been beyond the reach of AI. According to The Information, Strawberry can solve complex math problems it has never encountered, develop sophisticated marketing strategies, and even crack intricate word puzzles like the New York Times’ “Connections.” Impressively, Strawberry is said to have scored over 90% on the MATH benchmark, significantly surpassing the 53% scored by GPT-4 and the 76.6% achieved by GPT-4o. |
“OpenAI’s pursuit of human-level reasoning isn’t just a technological marvel; it’s a narrative of pushing boundaries and sparking new possibilities in every sector.” |
|
-Ghazenfer Mansoor, founder and CEO of Technology Rivers |
|
|
What’s the significance? As OpenAI prepares for a possible fall release, all eyes are on whether Strawberry will truly revolutionize AI capabilities and provide us with the big leap in AI many have been looking for. |
0 Responses
Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.