A report published November 9, 2026, said MiniMax, Moonshot AI and ByteDance Seed were recruiting for U.S.-based roles. The reported openings spanned San Francisco, San Jose, San Diego and Seattle; a separate account described a conditional Seed offer strategy tied to candidates who already had offers from OpenAI or Anthropic.

The U.S. roles the report described

Rest of World reported about 16 MiniMax vacancies in San Francisco, including a research-lead role focused on the company’s core AI model. It also said Moonshot AI was seeking U.S.-based solutions architects to help businesses integrate its models. ByteDance Seed, the company’s frontier-model division, was reported to have 71 positions listed across San Jose, San Diego and Seattle.

These are recruitment figures reported in the November 9 article. The report did not describe a count of completed hires.

What the Seed offer account says

A person familiar with ByteDance Seed told Rest of World that the division would try to make a competing offer described as “1.5 times higher” when a researcher already had an offer from OpenAI or Anthropic. That account describes a conditional recruiting approach, not a general salary benchmark for Seed or Chinese AI companies.

Why the pay figures do not make a salary ranking

The figures cited in the report cover different kinds of compensation and different groups of workers. MiniMax’s reported 2025 employee remuneration was $84.3 million across the company. Divided by its 415 employees at year’s end, that is about $203,000 per employee, including share-based compensation and staff in every department. It is not a researcher’s salary.

Z.AI’s reported $140,000 per employee was an annualized remuneration cost for the first half of 2026, also covering company-wide staff and including share-based compensation. OpenAI’s reported $250,000–$445,000 figure was an advertised annual base-salary range for research engineers in San Francisco, before equity awards. Those measures cannot be lined up as equivalent salaries.

Three Chinese researchers working in the United States told Rest of World that packages above $1 million a year were common among their peers. That was their account of peer compensation, not a standardized salary survey. Taken together, the figures do not establish that Chinese AI companies generally pay more than U.S. labs.

What researchers said matters beyond pay

The same three researchers said they preferred U.S. companies for work-life balance and the opportunity to work on frontier models. Their comments offer individual reasons for choosing an employer, rather than a measure of what researchers broadly prefer. Cheryl Wu, an AI-economics researcher at Yale University and GovAI, also identified compensation and career prospects as factors that can make recruitment difficult for Chinese companies.

A recruiting pitch and a company-wide pay average answer different questions; neither, on its own, tells you what a specific researcher was offered.