OpenAI Launches GPT-5.4 With Native Computer Use, 1M Token Context, and Record-Breaking Professional Task Performance

OpenAI launches GPT-5.4, a powerhouse AI model featuring native computer use, a 1-million token context window, and record-breaking 83% performance on professional tasks, surpassing human desktop navigation benchmarks while cutting false claims by 33% and token usage by 47%.

• OpenAI is releasing GPT-5.4, its most capable and efficient frontier model for professional work, available in ChatGPT as GPT-5.4 Thinking, the API, and Codex, with a GPT-5.4 Pro version also launching for maximum performance on complex tasks.
• GPT-5.4 delivers major advances across knowledge work, coding, and computer use, achieving 83% on GDPval professional tasks, 75% on OSWorld-Verified desktop navigation surpassing human performance, and matching GPT-5.3-Codex on SWE-Bench Pro while being faster and more token-efficient.
• The model introduces native computer-use capabilities, a 1M token context window, and a new tool search feature that reduces token usage by 47% in multi-tool workflows, while also being 33% less likely to produce false claims compared to GPT-5.2.