OpenAI’s launch strategy is more strategic than it appears: The dual launch of open-source models and GPT-5 serves different but complementary markets. The former allows companies to keep their data internally and fine-tune models for specific tasks. GPT-5, on the other hand, provides cutting-edge capabilities to the general population through the API and the ChatGPT application.
Open-source models solve the enterprise puzzle: “When you talk to demanding customers, like governments, banks, life sciences companies,” noted Godement, “the implementation of AI is going to be a hybrid between closed-source and open-source.”
We are witnessing the fastest technology adoption in recent memory: GPT-5 reached 2 billion tokens per minute in just 2 hours after launch, a scale that is “even hard to visualize for me, who spent the day in the infrastructure world.” Unlike traditional enterprise software migrations that take months, companies are changing in a matter of hours.
GPT-5 represents a true functional leap in AI capabilities: It may not have met AGI expectations, but the new model is still excellent for long-horizon tasks, staying on track for several minutes instead of just a few interactions. This allows for complex queries involving multiple systems and extended conversations.
Prices are dropping faster than expected: GPT-5 is 25 times cheaper than GPT-4 at launch, while being significantly more capable. Godement predicts the price will drop even further. For OpenAI, the goal is to make AI affordable enough to customize the experience for every user, “even if that homepage is viewed by billions of people.”