GPT-5 Stumbles With Consumers but Surges in Enterprise Adoption

The rollout of GPT-5 was met with mixed reactions from consumers, who found the model less intuitive than its predecessor, prompting OpenAI to revert paying users back to GPT-4. But behind the scenes, the new AI model is making significant inroads where it matters most: the enterprise market. While consumer-facing excitement fizzled, GPT-5 is quickly becoming the preferred choice for startups and businesses building AI-powered tools. Companies like Cursor, Vercel, and Factory have already made GPT-5 their default model in key products, citing faster setup times, superior performance on complex tasks, and lower costs. Vercel, a platform for building web applications, has integrated GPT-5 into its new "vibe coding" system, which transforms plain-English prompts into live, functional apps. The model has also been rolled out into its in-dashboard AI agent, where it excels at managing multi-step workflows and interpreting long, detailed instructions. Malte Ubl, Vercel’s CTO, said Claude had long dominated the coding space, but GPT-5 has now caught up — and in some areas, surpassed it, especially in early-stage prototyping and creative problem-solving. Pricing has been a major factor in this shift. Factory’s co-founder Grinberg noted that lower inference costs make users more willing to experiment freely, removing hesitation around asking follow-up questions or testing new ideas. “They can shoot from the hip more readily,” he said. AI-powered coding platform Qodo tested GPT-5 against top models including Gemini 2.5, Claude Sonnet 4, and Grok 4. It found GPT-5 led in detecting critical coding errors — such as security vulnerabilities and broken logic — often identifying issues that other models missed. While it occasionally produced false positives and redundant output, the overall performance was impressive. Box, a collaboration platform, has been testing GPT-5 on complex business documents like hundred-page lease agreements and product roadmaps. CEO Aaron Levie called the model a “breakthrough,” highlighting its ability to reason through long, logic-heavy tasks that earlier systems struggled with. For enterprise use, where AI agents automate workflows behind the scenes, these improvements are transformative. OpenAI has built a dedicated enterprise sales team of over 500 people under COO Brad Lightcap, operating independently of Microsoft, its longtime cloud partner. Customers can access GPT models via Azure or directly through OpenAI’s API, giving the company full control over the product experience. Despite the momentum, the enterprise AI race remains costly. Both OpenAI and Anthropic are investing heavily, with OpenAI on track to burn $8 billion this year. Still, OpenAI is gaining ground. GPT-5 API usage has surged, with coding and agent-building work more than doubling, and reasoning tasks increasing over eightfold since launch. Anthropic, meanwhile, remains a strong player with deep enterprise integration. Its Claude model powers tools for Amazon Prime, Alexa, and AIG, and is used across pharma, retail, aviation, and professional services. It’s embedded in AWS, GCP, Snowflake, Databricks, and Palantir. The company’s enterprise revenue has grown 17 times year-over-year, with $3 billion in revenue in just six months — including $1 billion in June alone. Average customer spend has more than quintupled, and over half of clients now use multiple Claude products. Still, OpenAI’s push into enterprise is accelerating. With GPT-5’s performance gains, pricing flexibility, and growing adoption by key players, the company is no longer just chasing consumers — it’s building a powerful foundation in the business world, where the real value lies.