Just weeks after its last major upgrade, OpenAI is shifting ChatGPT from a consumer-facing novelty to a core enterprise tool. On Thursday, the company unveiled GPT-5.2, a faster and more reliable model designed for complex professional workflows.

OpenAI said GPT-5.2 is now significantly better at handling tasks such as building spreadsheets and presentations, writing code, interpreting images, working with long context, calling tools, and executing multi-step projects—reflecting a broader push to embed its technology in daily business operations. The shift follows lucrative deals with both the U.S. government and Disney.

A new benchmark for workplace automation

OpenAI introduced a new proprietary benchmark called GDPval, which simulates tasks across 44 occupations. According to the company, GPT-5.2 matched or outperformed human workers in roughly 71% of comparisons.

Fidji Simo, OpenAI’s CEO of Applications, wrote on X that the “thinking” model beats or ties human experts in 70.9% of common professional tasks. However, it remains unclear whether GDPval has undergone independent external review.

Three models built for different needs

GPT-5.2 is available across all paid tiers, with API access live immediately. Developers can choose from three specialized variants:

Instant – optimized for fast, simple professional tasks

Thinking – for complex, multi-step workflows

Pro – the top-tier model for research and long-form work

API pricing is set at $1.75 per million input tokens and $14 per million output tokens.

OpenAI says GPT-5.2 also posted stronger results on existing benchmarks, including GPQA Diamond and FrontierMath, and shows notable improvements in coding, data analysis, and experimental design.