OpenAI's Q2 revenue isn't really the key point—the focus should be on Q3 revenue after the release of GPT-5.6. Today, the Wall Street Journal cited insiders reporting that OpenAI's Q2 revenue was $6.7 billion, an 18% increase from Q1's $5.7 billion. At the same time, operating losses further expanded, and everyone started panicking again. But let me just say this: OpenAI's weak performance in the first half of the year was obvious to the entire market. OpenAI's comeback started happening between late June and early July (end of Q2, start of Q3): GPT-5.6 began limited preview on June 26 and was officially released on July 9. The reputation and market response to GPT-5.6 showed a significant improvement compared to previous versions. Then, on July 30, OpenAI announced major price cuts: Luna dropped by 80% (input from $1 to $0.20, output from $6 to $1.20), Terra dropped by 20% (input from $2.50 to $2, output from $15 to $12), while Sol remained unchanged at $5 input / $30 output. They actively initiated a price war. So, OpenAI's weak Q2 performance was clear, and naturally, the revenue wasn't great either. What really needs attention is the Q3 revenue after GPT-5.6's release. As mentioned earlier here: https://((x.com))/qinbafrank/status/2082669379762950405?s=46&t=k6rimWsEbo2D2tXolYcM-A, OpenAI's revenue in July was actually catching up quickly. After all, in just two weeks in mid-July, Codex+GPTwork users grew by over 4 million. This was previously announced officially here: https://((x.com))/thsottiaux/status/2079609157934886975?s=46&t=k6rimWsEbo2D2tXolYcM-A.