According to CNBC, the AI darling now boasts nearly 700 million weekly users, but GPT-5’s debut prompted enough user whining that OpenAI had to quietly reinstate the older GPT-4 model for its paying chatbot punters. That said, GPT-5 was never about keeping casual users happy.

The real game is the enterprise market where rival Anthropic has enjoyed a head start. Startups like Cursor, Vercel and Factory have already made GPT-5 their default in key tools, pointing to its snappier setup, better handling of complex tasks and a much lower price.

Box CEO Aaron Levie reckons GPT-5 is a “breakthrough”, saying it chews through logic-heavy documents in a way older models simply could not. Companies working on code and UI design say GPT-5 is finally matching or beating Claude, Anthropic’s prized model, in areas it used to own outright.

The cost advantage is eye-watering. In some cases, GPT-5 is reportedly seven and a half times cheaper than Anthropic’s Claude Opus 4.1, and that kind of price cut turns heads even if the model occasionally dribbles out false positives or redundant output.

Qodo’s recent benchmark showed GPT-5 outperformed Google’s Gemini 2.5, Claude Sonnet 4 and Grok 4 in catching security flaws and code bugs, offering cleaner, more surgical fixes.

Even JetBrains has made GPT-5 the default for its AI Assistant and no-code tool Kineto, suggesting that its coding chops are holding up in real-world scenarios.

Despite the early user grumbles, GPT-5’s enterprise traction is soaring. One source told CNBC that the model is now handling more than double the coding and agent-building work since launch, with multi-step reasoning usage jumping eightfold. Businesses like the speed, the cost, and the fact it doesn’t hallucinate quite as often when you ask it to solve actual problems.

OpenAI is reportedly on track to burn through $8 billion this year just to keep the lights on and customers happy. Both it and Anthropic are deep in the VC begging bowl again, trying to scrape up the billions needed to fund this high-stakes arms race.

For OpenAI, the strategy is clear. Slash prices, eat the infrastructure bill, flood the zone, and hope that customer loyalty turns into something approximating profit before the next model drops.