OpenAI launches GPT-5.4 with Pro and Thinking versions | TechCrunch

"OpenAI released a GPT-5.4, a new foundation model billed as 'our most capable and efficient frontier model for professional work.' In addition to the standard version, GPT-5.4 is also available as a reasoning model (GPT-5.4 Thinking) or optimized for high performance (GPT-5.4 Pro). The API version of the model will be available with context windows as large as 1 million tokens, by far the largest context window available from OpenAI."

"OpenAI also emphasized improved token efficiency, saying GPT-5.4 was able to solve the same problems with significantly fewer tokens than its predecessor. The new model comes with significantly improved benchmark results, including record scores in computer use benchmarks OSWorld-Verified and WebArena Verified. The new model also scored a record 83 percent on OpenAI's GDPval test for knowledge work tasks."

"GPT-5.4 continues the company's efforts to limit hallucinations and factual errors. OpenAI said the new model was 33% less likely to make errors in individual claims when compared to GPT 5.2, and overall responses were 18% less likely to contain errors."

"[GPT-5.4] excels at creating long-horizon deliverables such as slide decks, financial models, and legal analysis, delivering top performance while running faster and at a lower cost than competitive frontier models."

OpenAI released GPT-5.4, positioning it as the most capable and efficient frontier model for professional work. The model is available in three variants: standard, reasoning (GPT-5.4 Thinking), and high-performance (GPT-5.4 Pro). The API version supports context windows up to 1 million tokens, OpenAI's largest offering. GPT-5.4 demonstrates significantly improved token efficiency, solving problems with fewer tokens than predecessors. The model achieved record benchmark scores including 83% on OpenAI's GDPval test and top performance on computer use benchmarks OSWorld-Verified and WebArena Verified. It excels at creating complex deliverables like slide decks, financial models, and legal analysis. Error rates decreased substantially: 33% fewer errors in individual claims compared to GPT-5.2, with overall responses 18% less likely to contain errors. A new Tool Search system optimizes API tool calling by allowing models to look up tool definitions as needed rather than loading all definitions upfront.

Read at TechCrunch