OpenAI’s long-awaited GPT-5 model, released on 7 August, has drawn mixed early reviews from consumers but is being adopted rapidly by corporate users. The company says API traffic doubled within the first 48 hours and internal data show a two-fold increase in coding and agent work alongside an eight-fold jump in reasoning tasks. Large enterprise runs highlight the model’s price-performance edge: one 30-million-token job cost $4.75, while the high-end version of GPT-5 delivered a 65.7 % ARC-AGI-1 score at roughly half the cost of rival Grok 4 and one-eighth that of Anthropic’s O3-Pro. Chief Executive Officer Sam Altman said the system was “shaped by Indian user feedback” and called India OpenAI’s second-largest—and fastest-growing—market, adding that it could soon become the company’s biggest. The public service already counts nearly 700 million weekly users worldwide. Altman nonetheless acknowledged the launch was “screwed up in parts,” noting that user expectations outpaced the incremental gains GPT-5 delivered and warning that OpenAI will need “billions” of dollars more in computing infrastructure to meet demand. In response, OpenAI has begun incremental fixes: GPT-5 now offers Auto, Fast and Thinking modes; ChatGPT Plus and Team subscribers receive up to 3,000 GPT-5 Thinking messages per week; and popular predecessor GPT-4o has been restored as a legacy option. Product managers say the new settings soften an overly formal tone and give customers clearer control over speed, depth and cost. Despite lingering criticism about benchmark scores and energy use, investors are betting on expanding revenue streams. French business daily La Presse and other outlets report that secondary-share transactions could lift OpenAI’s valuation from about $300 billion to as much as $500 billion, underscoring market confidence that enterprise subscriptions—and a surging user base in India—will outweigh the model’s rocky public debut.