OpenAI Update Gets 1M Context Window

OpenAI's latest model update includes 1M context window, native computer use, stronger reasoning, and faster responses — earning 109 likes and 11 reposts. The powerful update represents a significant leap in AI capabilities for enterprise and developer use cases.

The new model, officially named GPT-5.4, is OpenAI's first general-purpose AI with native computer-use capabilities. On the OSWorld-Verified benchmark for navigating desktop environments, GPT-5.4 achieved a 75% success rate, surpassing the human performance level of 72.4%. This "computer use" feature allows the AI to directly interact with software by analyzing screenshots and issuing mouse and keyboard commands. This enables the automation of complex workflows across multiple applications without manual intervention. The update expands the context window to 1 million tokens, a significant increase from the 400,000-token limit of its predecessor, GPT-5.2. This allows the model to process and analyze the equivalent of entire codebases or very long documents in a single prompt. This 1M token capacity brings OpenAI's model in line with long-context offerings from competitors like Google and Anthropic. While OpenAI's earlier GPT-4.1 model featured a 1M window, this capability was not present in the more recent GPT-5.2. Accuracy has also been improved, with GPT-5.4 being 33% less likely to make false individual claims and 18% less likely to produce a response containing any error compared to GPT-5.2. On a benchmark for professional knowledge work, its score jumped to 83.0% from GPT-5.2's 70.9%. The release is coupled with a push into finance workflows, including a new ChatGPT for Excel beta and improved performance on financial modeling tasks, where it scored 87.3% on an internal benchmark versus 68.4% for the prior model. For developers, a new API feature called "Tool Search" can reduce token consumption by as much as 47% in tasks that use many tools, lowering cost and latency. The model's vision capabilities were also upgraded to better understand high-resolution images, improving its accuracy in parsing documents and operating software. The update is being rolled out in several variants: GPT-5.4 Thinking is available for ChatGPT Plus, Team, and Pro users, while the higher-performance GPT-5.4 Pro is accessible via the API and for Enterprise customers.

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.