The Next Frontier in AI
OpenAI announced GPT-5, its most capable language model to date, at a live event attended by hundreds of developers and streamed to millions online. The model introduces a one-million-token context window — equivalent to roughly 750,000 words — enabling it to analyse entire codebases, legal documents, or books in a single prompt.
What's New in GPT-5
- Real-time web access: The model can browse the live internet to retrieve current information.
- Multi-step reasoning: Significantly improved performance on mathematical proofs, coding challenges, and scientific problem-solving.
- Multimodal input: Accepts text, images, audio, and video in a single prompt.
- Extended memory: Optional persistent memory across conversations.
"GPT-5 is not just an incremental update. It represents a qualitative leap in what AI systems can do," said OpenAI's CEO at the launch event.
Who Gets Access
GPT-5 rolls out first to ChatGPT Pro subscribers and API partners, with broader availability within 60 days. Pricing for the API is set at $15 per million input tokens — a 40% reduction compared to GPT-4 Turbo.
