<cite index="83-2">OpenAI is on track to generate annualized revenue of more than $40 billion based on its current performance, roughly doubling its run rate from the end of 2025</cite>. <cite index="86-1">OpenAI President Greg Brockman told staff the run rate rose more than 20% month-on-month in July</cite>. <cite index="87-2">Momentum was driven by the release of the company's GPT-5.6 series of models, its new enterprise agent ChatGPT Work, and growing adoption of its AI coding tool Codex</cite>.
<cite index="89-2">Anthropic generated more than $11.5 billion in preliminary revenue in Q2, more than 14 times what it generated in the same quarter last year</cite>, while <cite index="89-2">OpenAI's latest revenue run rate hit $40 billion per an internal message shared by co-founder Greg Brockman</cite>. <cite index="89-2">Anthropic's reported $65 billion run rate was hit at the end of July and implies a sevenfold increase from where it was at the end of last year</cite>—putting Anthropic ahead of OpenAI by valuation ($965B vs. $852B) and by ARR ($65B vs. $40B).
<cite index="88-2,88-3">OpenAI's annualized revenue has topped $40 billion, with growth coming as the price of AI is falling; prices for leading U.S. models have dropped by almost a quarter since mid-July amid intensifying competition from cheaper Chinese rivals such as DeepSeek and Moonshot</cite>. <cite index="88-4">OpenAI says improvements to its inference systems have reduced the end-to-end cost of serving GPT-5.6 by 20% and lifted token-generation efficiency by more than 15%</cite>.
For architects managing API costs and selecting between OpenAI and Anthropic: the 20%+ July growth in OpenAI's ARR reflects strong consumption velocity in enterprise and coding. But Anthropic's margin advantage (54% of Claude Code revenue is enterprise at higher per-token cost) is now visible in run rate. OpenAI's 20% cost reduction on GPT-5.6 and price cuts on token costs may be necessary for share defense, but tokenomics alone no longer favor OpenAI. The divergence in enterprise attachment and gross-margin structures suggests 2H2026 will be dominated by API pricing warfare, not just model capability comparisons.