OpenAI's O1 AI Models: Expanding Reach and Advancing AI Capabilities

OpenAI expands o1 AI models to enterprise and education, competing directly with Anthropic

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More OpenAI has made its latest AI models, o1-preview and o1-mini, available to all ChatGPT Enterprise and ChatGPT Edu customers. These models, designed to handle complex reasoning tasks, are poised to change how organizations and academic institutions tackle their most difficult challenges, from advanced coding to scientific research. The o1 models, first announced earlier this month, represent OpenAI's most advanced attempt yet at creating AI capable of deep, multi-step reasoning. By imitating human thought processes, these models can solve intricate problems that earlier AI iterations struggled with, offering new possibilities for industries reliant on advanced problem-solving. AI designed to think: What makes o1 models different The o1-preview and o1-mini models are built to think more critically and deeply than their predecessors. OpenAI trained these models to spend more time processing information before responding, allowing them to handle complex tasks in areas like mathematics, coding, and scientific discovery. In early tests, o1-preview demonstrated its capabilities by solving 83% of problems in a qualifying exam for the International Mathematics Olympiad -- a substantial improvement over GPT-4o, which managed only 13%. Similarly, the model excelled in coding competitions, ranking in the 89th percentile on Codeforces, a platform where coding skills are rigorously tested. The smaller, more cost-efficient o1-mini model is tailored specifically for coding tasks, offering a more affordable option for companies that need advanced problem-solving without the need for broad world knowledge. This makes o1-mini particularly useful for tasks like generating and debugging complex code, providing an accessible option for smaller businesses and developers. Why o1 models are a game-changer for enterprises For enterprise customers, the new o1 models represent a significant leap forward. Businesses across industries -- from finance to healthcare -- are increasingly turning to AI not just for automation but to solve intricate, high-stakes problems where human expertise is limited. The o1 models' ability to reason, refine strategies, and recognize mistakes makes them ideal for these use cases. These capabilities are particularly attractive for companies dealing with complex data sets and workflows. The o1-preview model, for example, can assist physicists in generating complex quantum optics formulas or help healthcare researchers annotate large-scale genomic data. This is a stark contrast from earlier AI models that primarily handled repetitive, low-level tasks. Dr. Derya Unutmaz, an immunologist at The Jackson Laboratory, recently used the o1-preview model to write a cancer treatment proposal. "It created the full framework of the project in under a minute, with highly creative aims and even considerations for potential pitfalls," he posted on X.com (formerly Twitter). "This would have taken me days, if not longer, to prepare," he added, noting that the model brought up ideas he might not have considered himself, even with 30 years of experience in the field. This kind of productivity and creativity boost is why so many businesses are eager to integrate these models into their workflows. OpenAI's decision to prioritize enterprise customers with this release highlights its strategy to capture the high-value, high-complexity segment of the AI market. Educational institutions stand to benefit immensely The o1 models are also a powerful tool for educational institutions. Universities and research centers often face resource and time constraints when conducting complex data analysis or research. By making these models available to ChatGPT Edu customers, OpenAI is giving students and researchers access to cutting-edge AI tools that can help them tackle some of the most difficult problems in their respective fields. Initial feedback from the academic community has been overwhelmingly positive. Dr. Kyle Kabasares, an astrophysicist at the Bay Area Environmental Research Institute, posted on X.com that o1-preview "accomplished in 1 hour what took me about a year during my PhD." In fields like computational fluid dynamics and immunology, where complex calculations and data analysis are routine, the o1 models have already proven their value by speeding up research processes and offering new insights. The o1 models are also poised to change how students learn. By handling more complex tasks, these models allow students to focus on higher-level thinking rather than getting bogged down in rote processes. This shift could lead to more innovation and creativity in academic research, accelerating breakthroughs in fields ranging from physics to biology. Safety and governance: OpenAI's commitment to responsible AI In addition to their advanced capabilities, the o1 models come with enhanced safety features. OpenAI has developed a new safety training approach that allows these models to reason through ethical guidelines and safety rules. This is crucial for enterprises and educational institutions handling sensitive data. OpenAI has stated that it does not use customer data for training, ensuring that proprietary information remains secure. The company has also introduced rigorous safety evaluations, including a test known as "jailbreaking," where o1-preview scored 84 out of 100, far surpassing GPT-4o's score of 22. This means the o1 models are better equipped to resist attempts to bypass safety protocols, a critical feature for businesses concerned about compliance and data privacy. In a broader context, OpenAI has formalized partnerships with AI safety institutes in the U.S. and U.K., giving these organizations early access to the models for independent testing. This collaboration aims to ensure that AI advancements are aligned with ethical guidelines and regulatory frameworks, a growing concern as AI systems become more autonomous and integrated into daily operations. The competitive landscape: OpenAI vs. Anthropic The release of the o1 models positions OpenAI as a leader in the highly competitive AI enterprise space. However, the company faces strong competition. Anthropic, another major player in AI, recently launched its own enterprise-focused model, Claude Enterprise, which offers a massive 500,000-token context window -- more than double what OpenAI's models currently provide. While Anthropic's models excel in processing large data sets, OpenAI's strength lies in its focus on deep reasoning and problem-solving. OpenAI's ability to integrate these advanced models into its existing enterprise and educational offerings gives it a competitive edge. While Anthropic may have the upper hand in data processing capacity, OpenAI's focus on reasoning tasks could give it a long-term advantage, especially in industries where problem-solving is more valuable than sheer data crunching. The future of AI in business and education The introduction of OpenAI's o1-preview and o1-mini models signals a turning point in the landscape of artificial intelligence. These models go beyond automating routine tasks -- they're designed to think critically, making them true partners in tackling the toughest challenges in industries like healthcare, quantum research, and advanced coding. As businesses and educational institutions increasingly rely on AI for high-stakes decision-making and complex problem-solving, the impact of these models could reshape what we expect from intelligent systems. In a world where innovation often happens at the intersection of technology and human insight, the o1 series offers a bridge to the future. It's no longer about what AI can do -- it's about what AI should do. And with OpenAI's latest leap forward, the answer seems clear: it should do a lot more.

OpenAI's o1: More Than Meets The Eye

With all the buzz around OpenAI's project "Strawberry," I was eager to try out OpenAI's o1 preview when it launched. At first, it felt like an incremental update. The more I explored, the more I realized this model is a significant step forward and a preview of what is to come. Here's why: I had hoped OpenAI would implement "self-taught reasoning," where models can evaluate and refine their internal processing (something akin to human "thoughts"). While o1 isn't there yet, it combines three key innovations: Deep Reinforcement Learning (Q-learning), "Chain of Thought" (CoT), and a "Tree of Thoughts" approach: OpenAI improved the model's ability to reason through safety protocols, making it much more resistant to jailbreak attempts (efforts to bypass its safeguards). In safety tests, o1 scored 84 out of 100, compared to GPT-4's score of just 22. OpenAI is also working with AI safety institutes in the U.S. and U.K. to further evaluate and refine these capabilities. This improvement will make o1 a strong candidate for future applications where AI agents must operate autonomously while adhering to company policies and regulations. In its current form a lack of tool access that prevent the model preview from taking actions. If you want to do a fun demo, ask Chat GPT 4o how many r's are in the word Strawberry. It may tell you 2. This is because the model represents the word as tokens rather than letter by letter. Ask o1 and you will see it think for a split second and get the answer right. To test both model's capabilities, I asked both ChatGPT 4o and o1 to develop a quantum circuit that solves a Max-Cut optimization problem. o1 clearly outperformed GPT-4, not only delivering a better solution than GPT 4o, but also providing a detailed explanation of its reasoning process. This transparency is crucial for business applications in regulated industries, where explainability is key. The additional accuracy comes at the cost of time - o1 takes longer to generate results. In my case, o1 took 8 seconds more than GPT-4. This makes it unsuitable for real-time applications, but ideal for decision-support systems where detailed reasoning is more important than speed. The model's higher computational demands also translate into a higher price: $15 per 1 million input tokens and $60 per 1 million output tokens, compared to GPT-4o's $5 and $15, respectively. Also, you pay for the tokens it uses in internal "thinking" as well as tokens for input and output. Businesses will need to weigh o1's capabilities against its cost and determine where it fits into their system architecture. At first glance, o1 may seem like a minor update, but it marks a major step forward in AI reasoning. As OpenAI's strategy of steady improvements released incrementally continues, improvements in problem-solving, explainability, and safety lay the groundwork for future breakthroughs. I hope introspection and self-teaching are coming soon. While the higher cost and slower speed are trade-offs, o1 is better for use cases where transparency and accuracy are essential and can justify the extra resources. As you think through what o1 means for your generative and agentic AI aspirations, clients can have a guidance session with me to discuss what this all means in the short and long term and how you plan for the rapid pace of AI progress.

OpenAI o1 Likely Uses RL over Chains of Thought to Build System 2 LLMs

OpenAI's o1 could be considered the first successful commercial launch of a System 2 LLM. Recently, OpenAI released two models - OpenAI o1-preview and OpenAI o1-mini - marking a significant leap in the AI world. These models can now reason using chain of thoughts and reasoning tokens. Jim Fan, in a recent post on X, mentioned that o1 models mark a significant shift towards inference-time scaling in AI, emphasising the importance of search and reasoning over mere knowledge accumulation. This approach suggests that effective reasoning can be achieved with smaller models. By implementing techniques like Monte Carlo tree search during inference, the model can explore multiple strategies and scenarios to converge on optimal solutions. The key advantage of using MCTS during inference is that it allows the model to consider many different approaches to a problem, rather than committing to a single strategy early on. Subbarao Kambhampati, professor at Arizona State University, said that OpenAI's o1 model uses reinforcement learning over auto-generated chain of thought -- similar to AlphaGo's self-play approach -- to optimise problem-solving by building a generalised System 2 component atop LLM substrates, albeit without guarantees. "One interesting issue with o1 is that it seems to be significantly less steerable compared to LLMs. For example, it often completely ignores any output formatting instructions making it hard to automatically check its solutions," he added, saying that once you are an approximate reasoner, you might develop the 'don't tell me how to solve the problem; I already have a way I use to solve it' complex. In his 2011 book Thinking, Fast and Slow, Daniel Kahneman coined the term 'System 2 thinking' which refers to complex problem-solving, logical reasoning, and careful decision-making, often involving step-by-step analysis and focused attention. Sounds very similar to what OpenAI has promised with its latest o1 model, which "thinks". Maybe, OpenAI's o1 could be considered the first successful commercial launch of a System 2 LLM and the most important reason for that is the reasoning tokens. These tokens are designed to guide the system to perform step-by-step reasoning. They are created based on the user's prompt and added to the reasoning process. Reasoning tokens in the systems are often notated with single or double-angle brackets for illustrative purposes. OpenAI decided to use English words as reasoning tokens for convenience, such as "Interesting", "First", "Let's test this theory", "Wait", "That seems unlikely", "Alternatively", "So this works", "Perfect", etc. With the use of reasoning tokens, o1 models demonstrated significantly better performance on complex tasks compared to previous models. For example, o1 solved 83% of problems in a qualifying exam for the International Mathematics Olympiad, compared to GPT-4's 13%. When AIM first tapped into ChatGPT's o1, our debut question was, "How many 'R's does 'Strawberry' have?" - and it nailed it. Later, we also asked which was bigger - 9.9 or 9.11 - and it got that right as well. This shows that OpenAI has finally solved Jagged Intelligence. Anatoly Geyfman, the co-founder and CEO of Carevoyance, explained that reasoning tokens are meant to be used for the time a model spends "thinking". It's used to pay for the additional submissions to itself to refine the process of arriving at an answer, or whatever the actual mechanism of action is. "This is important - there is now a way for model builders to monetise the more sophisticated actions of a model beyond "input" and "output" tokens. The reasoning tokens let OpenAI and, I bet, others in the near future release models that aren't so much better trained, but instead, are better at thinking through responses," he added further. A similar approach was mentioned in a paper titled 'Guiding Language Model Reasoning with Planning Tokens', published in July. It proposed adding specialised planning tokens at the beginning of each chain-of-thought step to guide and improve language models' maths reasoning ability. Saurabh Sarkar, the CEO of Phenx Machine Learning Technologies, mentioned that when you try to solve a question like "What is 2 + 2, then multiply the result by 3?" with a traditional approach, it will first calculate 2 + 2, get the result 4, and then multiply 4 by 3 to get 12. Using reasoning tokens, the model anticipates the need to multiply the intermediate result (4) with 3 while still calculating 2 + 2. It "pre-computes" and stores this information, so when it reaches the multiplication step, it already has the necessary data, allowing for faster and more efficient processing. This is how reasoning tokens allow for more thorough and accurate responses to challenging queries. Theo Browne, a popular YouTuber, founder and CEO of Ping Labs, recently posted a video on the reasoning capabilities of o1 models. In response to the popular saying "We have PhD in our pocket," Browne said, "PhD that can't do basic maths" as the o1 models were not able to find all the possible corners of a parallelogram. https://x.com/allgarbled/status/1834344480797057307 (Please embed this in HTML) A Reddit user mentioned that OpenAI is advertising this as some kind of mega-assistant for research scientists and quantum physicists. "I gave it a fairly simple twin paradox time dilation problem, and it failed just as miserably as all the previous versions. It seems like it still has no understanding, just probabilistic word guessing," he added. He suggested that even after using reasoning tokens and taking more time to generate the answer, the model does not give satisfactory results. Another user mentioned that o1 was, in fact, performing worse than ChatGPT 4o. He mentioned that the responses from o1 were wordy, generic and 'safe' and he had to coax it several times to give him the same response that GPT4 provided on the first try. Apart from the reasoning capabilities of reasoning tokens, not showing tokens to API users raised concerns amongst users.