The latest OpenAI models, 03 and 04 Mini, have demonstrated remarkable intelligence and reasoning capabilities. The 03 model has achieved the highest IQ score on the planet, scoring 136 on the Mensa IQ test, surpassing the previous record holder, Gemini 2.5 Pro, which scored 128.

One of the most impressive features of the 03 model is its ability to use tools effectively and iteratively during its chain of thought. This allows it to handle complex, multi-step tasks with incredible precision and generate insightful scientific hypotheses on demand. The model's responses on medical and clinical questions sound like they are coming directly from top specialist physicians, showcasing its remarkable expertise.

The 04 Mini model has also impressed with its capabilities, particularly in the areas of tool usage within the reasoning chain and problem-solving. It has demonstrated the ability to solve challenging math problems, such as the latest Project Euler problem, far faster than any human solver. Additionally, the model has excelled in coding tasks, achieving the top spot on the coding intelligence index.

These models have set new benchmarks for AI capabilities, with their performance on tasks like geoging, maze-solving, and coding demonstrating their versatility and problem-solving prowess. While they are not perfect and may still fail on certain tests, the overall progress in AI technology showcased by these models is truly remarkable and represents a significant milestone in the field.

The recent release of OpenAI's 03 and 04 Mini models has sparked a strong reaction in the industry, with experts highlighting their remarkable capabilities. These models have demonstrated an unprecedented level of intelligence, solving complex tasks with ease.

One of the most impressive feats of the 03 model is its ability to excel at geoging, a task that requires identifying the location of a random Google Street View screenshot. The model was able to accurately pinpoint the location, even in challenging scenarios, surpassing the performance of expert human geogessers.

Furthermore, the 04 Mini model has showcased its prowess in solving advanced math problems. It was able to solve a recent Project Euler problem in just 2 minutes and 55 seconds, far outpacing the fastest human solvers, who took over 30 minutes to complete the task.

The models' coding abilities are also noteworthy. They have demonstrated flawless performance on challenging coding tasks, such as the hexagon and balls inside test, outperforming even the highly capable Gemini 2.5 Pro model.

These achievements highlight the remarkable progress in artificial intelligence, with the 03 and 04 Mini models setting new benchmarks for intelligence and problem-solving capabilities. As the industry continues to push the boundaries of what is possible, these models serve as a testament to the incredible potential of AI technology.

The recent releases of OpenAI's 03 and 04 Mini models have generated significant excitement in the AI industry. These models have demonstrated impressive capabilities that surpass many of their predecessors.

One of the key highlights is the 03 model's performance on the Mensa IQ test, where it achieved a score of 136, making it the highest IQ model on the planet. This exceeds the previous record holder, Gemini 2.5 Pro, which scored 128 on the IQ scale.

Another notable aspect is the models' ability to use tools effectively within their chain of thought. This allows them to tackle multi-step tasks with incredible reasoning and precision, generating complex and insightful scientific hypotheses on demand. Experts have likened the models' performance to that of top-level specialists in various domains.

Comparisons to other industry-leading models, such as Gemini 2.5 Pro and Deepseek R1, have also been made. The 03 and 04 Mini models have demonstrated superior performance on tasks like the "balls in hexagons" test, where they were able to seamlessly simulate the physics of the moving balls, while other models struggled.

Furthermore, the 04 Mini model has claimed the highest artificial analysis intelligence index score to date, showcasing its exceptional capabilities in areas like coding intelligence, where it has surpassed even the highly capable Gemini 2.5 Pro.

While the models are not without their limitations, as evidenced by their occasional failures on certain tests, the overall consensus is that the 03 and 04 Mini releases represent a significant milestone in the advancement of AI technology, with their impressive reasoning abilities and problem-solving skills.

The pricing of the new OpenAI models, 04 Mini and Gemini 2.5 Flash, is an important factor to consider. 04 Mini is priced in line with 03 Mini, though the cash tokens are half the price of 03 Mini. Meanwhile, the newly released Gemini 2.5 Flash is even cheaper than 04 Mini.

One of the key differences between these models is their context window size. 04 Mini has the same 200K token context window as 03 Mini, which is notably smaller than the massive 1 million token context window of Llama 4.1. Gemini 2.5 Pro also has a very large context window.

This smaller context window for 04 Mini means that as a reasoning model, it uses a high amount of tokens compared to other models, though it is marginally lower than 03 Mini. The ability to use fewer tokens in the thinking and chain of thought process is important, as it can lead to cheaper, faster, and more efficient performance.

Overall, the pricing and context window considerations are important factors to weigh when choosing between these powerful AI models and their capabilities.

The recent release of OpenAI's 03 and 04 Mini models has generated significant excitement and reactions within the AI industry. These models have demonstrated remarkable capabilities, surpassing previous benchmarks and showcasing impressive advancements.

The 03 model has been praised for its "genius-level" performance, achieving the highest IQ score on the Mensa test among AI models. Its ability to use tools effectively and iteratively during its chain of thought process has been particularly impressive. The model's reliability, precision, and capacity for generating complex, insightful, and evidence-based responses have been highlighted as standout features.

The 04 Mini model has also received praise for its impressive capabilities, including the ability to perform tool calls within its reasoning chain. This feature has been described as a significant unlock in the field of AI. The model's performance on various tasks, such as solving challenging math problems and coding challenges, has been nothing short of remarkable, often outperforming even the most skilled human experts.

While these models are not without their limitations, the overall consensus is that the advancements demonstrated by 03 and 04 Mini represent a significant milestone in the development of AI technology. The industry is eagerly anticipating the continued progress and potential applications of these groundbreaking models.