OpenAI has officially released GPT-5, featuring native multimodal processing across text, images, audio, and video, along with a context window of 1 million tokens. The model represents the most significant capability jump since GPT-4's release in 2023.
GPT-5's multimodal capabilities are truly integrated rather than bolted on. The model can analyze a video, read accompanying documents, listen to audio commentary, and synthesize insights across all modalities in a single reasoning chain. This enables applications that were previously impossible.
Benchmark performance is unprecedented. GPT-5 achieves expert-level scores on professional exams in law, medicine, engineering, and finance. On complex reasoning tasks, it matches or exceeds the performance of domain specialists in controlled studies.
The 1 million token context window means the model can process entire books, lengthy legal documents, or hours of video in a single session. This dramatically expands the scope of tasks AI can handle without losing context or coherence.
Pricing for API access starts at $15 per million input tokens and $60 per million output tokens, a premium over GPT-4 but competitive given the capability upgrade. ChatGPT Plus subscribers get limited GPT-5 access at the existing $20/month price.
