OpenAI reported that Codex, its agentic AI tool, has become the main work tool across the company, with 97.9% of active users and 99.8% of output tokens generated through the platform as of June 2026. The company published an economic research paper on Thursday that tracked adoption patterns for Codex from public release through mid-2026.
Adoption among engineers was gradual initially, with the average engineer shifting the majority of their usage to Codex by December 2025. Non-technical departments - Legal, Finance and Recruiting - reached majority Codex use later, around April 2026. The analysis indicates that the average lawyer or recruiter now produces more than 85% of their output tokens via Codex, while engineers produce 99% of their tokens on the platform.
Usage intensity also rose markedly. By May 2026, 80.6% of sampled individual users had made at least one Codex request estimated to replace more than 30 minutes of human work. Some 70.2% of those sampled made at least one request estimated to replace more than one hour of work, and 25.6% made at least one request estimated to replace more than eight hours of work. By June 2026, users at the 99th percentile were generating in excess of 60 hours of Codex agent turns per day on a regular basis.
Non-developer adoption outpaced developer adoption across all measured groups. Since August 2025, non-developer individual users increased 137-fold, non-developer organizational users rose 189-fold, and non-developer adoption within OpenAI itself expanded 12-fold. Across organizational users, Codex accounts for 63.3% of output tokens, while individual users generate 16.5% of their tokens via the platform.
The research departments exhibited the largest relative increase in output tokens, with median tokens rising 56 times between November 2025 and June 2026. Customer Support increased its median tokens 32-fold, Engineering 27-fold, and Legal 13-fold over the same interval.
The report included market ticker references in its presentation, listing MSFT, GOOGL and GOOG.
Methodology note - The findings summarized above are drawn from the company's internal economic research tracking Codex adoption since its public release; the paper provides the detailed counts and timing cited here.