OpenAI launches o1 model that can solve 'all maths, science problems,' locks it behind $200/month paywall

OpenAI o1 isn’t the GPT-5 many users are eagerly awaiting. However, it’s a significant step forward for AI-powered tools, especially in the face of growing competition from Google’s Gemini and Meta’s AI projects OpenAI has officially rolled out its highly anticipated o1 model for ChatGPT, promising a new level of problem-solving prowess. OpenAI has shared that the model scores an impressive 83 per cent on the notoriously difficult International Mathematics Olympiad qualifying exam — a far cry from the 13 per cent achieved by its predecessor, GPT-4o. OpenAI is also softening the blow with a grant programme to provide free access to ChatGPT Pro for medical researchers, underscoring the model’s potential for high-stakes applications. A step forward, but not GPT-5 Despite the excitement, OpenAI o1 isn’t the GPT-5 many users are eagerly awaiting.