DIGITIMES has noted that, in the absence of breakthrough technologies, GPT-5 has primarily focused on optimizing existing features. Key enhancements include expanding the context window to 400,000 tokens, improving reasoning and code generation, reducing hallucination rates, and supporting multimodal processing across text, images, and audio.
The introduction of a "Thinking" mode enables multi-step reasoning and long-form content generation, enhancing accuracy and reliability for complex tasks and making the model more suitable for academic and professional applications.
From a business perspective, GPT-5 has strengthened its role in education, with personalized learning modes that support customizable materials and long-form content. Its low hallucination rate ensures academic reliability, particularly benefiting science, technology, engineering, and mathematics (STEM) education and administrative integration, while also driving revenue growth for OpenAI's enterprise customers.
GPT-5 focuses on enhancing features, limited in tech breakthroughs
Enhanced GPT-5 features benefit & expand education applications
Chart 2: OpenAI GPT-5 enhanced features and major education applications
GPT-5 unifies legacy models into single one, saving users troubles
Chart 4: Context window and max. output length of major LLMs and features of GPT-5
GPT-5 focuses on code gen accuracy andreducing hallucination
Chart 5: Accuracy evaluation and hallucination rate of GPT-5 and o3 models
GPT-5 adds customizable window colors and personality traits
Chart 6: GPT-5 customizable window color and default personality traits
GPT-5 focuses on medical care knowledge & cut hallucination rate
Chart 7: Evaluation of GPT-5, o3, and GPT-4o in medical and health care fields
GPT-5 expands scope of knowledge and delivers expert-level responses
Chart 8: Evaluation of GPT-5, o3, and GPT-4o in knowledge fields
Chart 9: OpenAI revenues and losses and revenue structure, 2023-2024 (US$b)
Chart 10: Benefits of GPT-5 in OpenAI business model for education field
