OpenAI launched GPT-5, a system of models that are the company's most advanced with the promise of advanced reasoning, coding abilities and agentic features.
CEO Sam Altman said GPT-5 is like having "superpower on demand" and critical for enterprises and developers. GPT-5 will become the default across OpenAI services.
"GPT-5 is great for a lot of things, but we think it's going to be an especially important moment for businesses and developers, and we're very excited to see what they're going to build with this new technology.
For OpenAI, the stakes are huge as large language models (LLMs) take performance leaps almost daily. "GPT-5 is like talking to an expert, a legitimate PhD level expert in anything, about any area you need, on demand, which can help you with whatever your goals are," said Altman.
OpenAI launched the following flavors of GPT-5. As noted in its system card, OpenAI outlined the following progression.
- GPT-4o is replaced by gpt-5-main
- GPT-4o-mini replaced by gpt-5-main-mini
- OpenAI o3 replaced by gpt-5-thinking
- OpenAI o4-mini replaced by gpt-5-thinking-mini
- GPT-4.1-nano replaced by gpt-5-thinking-nano
- OpenAI o3 Pro replaced by gpt-5-thinking-pro
OpenAI said that GPT-5 is faster, more reliable and accurate on multiple tasks, including health.
The company said GPT-5 will roll out on all tiers--including free--and launch for enterprises and education next week. OpenAI plans to deprecate its previous models.
According to the GPT-5 system card:
"GPT‑5 is a unified system with a smart and fast model that answers most questions, a deeper reasoning model for harder problems, and a real-time router that quickly decides which model to use based on conversation type, complexity, tool needs, and explicit intent (for example, if you say “think hard about this” in the prompt). The router is continuously trained on real signals, including when users switch models, preference rates for responses, and measured correctness, improving over time. Once usage limits are reached, a mini version of each model handles remaining queries. In the near future, we plan to integrate these capabilities into a single model."
It remains to be seen whether GPT-5 moves the needle for Open-AI’s enterprise share. According to Menlo Ventures, Anthropic has 32% share of enterprise AI followed by OpenAI at 25% and Google at 20%. OpenAI spent a lot of time talking about GPT-5's ability to address health care use cases.
Microsoft said GPT-5 is now included into its consumer, developer and enterprise products including Microsoft Copilot, Microsoft 365 Copilot, Copilot Studio, Copilot Studio, GitHub Copilot, Visual Studio Code and Azure AI Foundry.
Key points include:
- GPT-5 is priced at $1.25/1M input tokens and $10/1M output tokens.
- GPT-5 mini is $0.25/1M input tokens and $2/1M output tokens.
- GPT-5 nano is $0.05/1M output tokens and $0.40/1M output tokens.
- OpenAI is offering four present personalities for GPT-5 including Cynic, Robot, Listener and Nerd.
- The API now has a verbostity setting for answers. I've been guilty of calling GPT a wordy 7th grader.
