OpenAI has stunned the tech world for the second time in a week, unveiling its newest flagship model, GPT-5, just days after releasing two open-source AI models. CEO Sam Altman boldly called GPT-5 “the best model in the world,” though early benchmarking shows it only slightly outperforms rival systems from Anthropic, Google DeepMind, and xAI in some areas and slightly underperforms in others.
Where GPT-5 is clearly leading is price. The API is set at $1.25 per million input tokens and $10 per million output tokens (plus $0.125 per million cached input tokens), matching Google’s Gemini 2.5 Pro basic rate but undercutting Anthropic’s Claude Opus 4.1, which starts at $15 for inputs and $75 for outputs. This competitive pricing, coupled with strong performance in coding tasks, has already earned praise from developers and sparked speculation about a looming large language model (LLM) price war.
Some observers believe the move could pressure competitors to slash their own rates, potentially reshaping the economics for AI-powered tools and startups that rely on these models. However, given massive AI infrastructure spending with OpenAI’s $30B Oracle deal, Meta’s projected $72B AI budget for 2025, and Alphabet’s $85B allocation, industry-wide price drops may still be limited.
Still, by dropping GPT-5 and open-source models in the same week, OpenAI has made it clear: the fight for AI dominance is as much about affordability as it is about capability.
