OpenAI launches AI models with humanlike ‘reasoning abilities’

• Looks like AI is ready to go head-to-head with human-level reasoning.
• OpenAI just unveiled its groundbreaking AI models; o1, and o1-mini.
• These models can not just answer competitive programming questions but even outperform human PhD-level benchmarks in science.

OpenAI’s latest series of AI models

OpenAI, backed by Microsoft, just dropped its latest series of AI models, code-named “Strawberry,”

Designed to process queries more thoroughly and tackle the hardest problems out there.

These new models, o1 and o1-mini, aren’t just faster; they’re smarter, especially in complex fields like science, coding, and math.

• o1 is a powerhouse in problem-solving, scoring an impressive 83% on the International Mathematics Olympiad qualifier, compared to GPT-4o’s 13%. It’s excelling in areas where AI traditionally struggled, like competitive programming, math, and even PhD-level science.
• o1-mini is a faster, cheaper version—80% less expensive—designed to handle tough STEM questions and debug complex code. It’s a cost-effective solution for tasks that require reasoning but don’t need extensive world knowledge.

What’s so special about o1 and o1-mini?

Their ability to think like humans—but faster.

Using a “chain-of-thought” reasoning technique, the model breaks down complex problems into smaller, manageable steps—much like a human would.

These models are trained to slow down and think through problems before responding, refining their approach as they go.

But unlike a human, it can do this faster and without user prompts.

Also read: Accenture’s $3Bn AI gamble pays off: CEO Announces Soaring AI Earnings and New GenAI Deals worth $900 Mn

Use cases

For instance, healthcare researchers can leverage o1 to annotate cell sequencing data.

Physicists can use it to generate complex mathematical formulas essential for quantum optics.

Developers across various fields can harness it to create and execute multi-step workflows.

Who will get access to OpenAI o1?

Starting 12th September, ChatGPT Plus and Team users got an exclusive sneak peek at the new o1 model.

Just select ‘o1-preview’ in the model selector, and you’re off to the races!

Next week, ChatGPT Enterprise and Education users will get their hands on both models.

Also read: Google lays off finance team amid AI focus. Operations shift to Bengaluru, Mexico, Dublin

Should finance professionals be worried?

The launch of new AI models with enhanced reasoning abilities, like those developed by OpenAI, may raise both opportunities and concerns for finance professionals.

While some may worry about job displacement, the real challenge lies in adaptation. Finance professionals who embrace AI tools like o1 will likely thrive.

AI can handle data-heavy tasks, but humans are still needed for decisions that require judgment, creativity, and emotional intelligence.