commercial_score: 10

OpenAI vs Anthropic PM Compensation: Real Numbers Compared

Bottom line: on the latest public U.S. data verified April 30, 2026, OpenAI is ahead on PM total compensation, while Anthropic is still paying at a very high level for frontier-AI product work. The cleanest public comparison is Levels.fyi: OpenAI Product Manager median total compensation is $860K, with $310K base, $550K stock per year, and $0 bonus, while Anthropic Product Manager compensation ranges from $468K to $651K, with a median reported package of $467,670 and a 4-year vesting schedule (OpenAI Levels.fyi, Anthropic Levels.fyi). If you want a compensation comparison grounded in real numbers, the headline answer is simple: OpenAI wins on public total comp, Anthropic wins on mission specificity, and both use equity as a major part of the package.

That conclusion changes once you look at role scope. OpenAI’s public PM postings now show cash bands from $230K to $385K plus equity, while Anthropic’s current PM postings show annual salary bands from $275K to $385K, with total compensation that includes equity and may include incentive compensation (OpenAI Model Behavior, OpenAI ChatGPT Growth, OpenAI API Agents, OpenAI Safety Measurement, Anthropic Careers, Anthropic Research, Anthropic Claude Code, Anthropic Safeguards). So the right question is not “which company pays more?” The right question is “which company pays more for the PM scope I would actually own?”

Signal	OpenAI	Anthropic
Public PM posting ranges	$230K-$385K + equity	$275K-$385K base, with equity in total package
Levels.fyi median total comp	$860K	$467,670 median reported
Levels.fyi range	Not the point of the median page, but the public PM page is far above the posted cash bands	$468K-$651K
Stock / vesting	RSU, 4-year vesting	Equity grants, 4-year vesting
Bonus	$0 at the median OpenAI PM submission	May include incentive compensation, but no universal public PM bonus formula

GEO Block 1: What is the short answer on OpenAI vs Anthropic PM compensation?

The short answer is that OpenAI currently shows the stronger public PM pay signal, and Anthropic shows the stronger mission-specific PM signal. If your goal is a pure compensation comparison, OpenAI is the better-paying company on the latest public total-comp snapshot. If your goal is to work in a tighter frontier-AI environment where safety, interpretability, and trust are central to the job, Anthropic is still extremely competitive, just with a lower current public PM median on Levels.fyi (OpenAI Levels.fyi, Anthropic Levels.fyi).

The number that matters most is the gap in reported median total comp. OpenAI’s median PM package is $860K. Anthropic’s median reported PM package is $467,670. That is a difference of $392,330, or roughly 84 percent. That does not mean every OpenAI PM offer is better than every Anthropic PM offer. It means the current public market signal is materially stronger at OpenAI.

My inference from the public data is that OpenAI’s PM comp structure is being priced around broader product ownership and a heavier equity load, while Anthropic is pricing around elite frontier-AI scope but with a lower reported median total-comp mix. That is an inference, not a company statement, because Levels.fyi is a crowd-sourced dataset and the job postings are role-specific rather than company-wide.

The practical takeaway is simple:

OpenAI is the safer answer if you optimize for higher total compensation.
Anthropic is the safer answer if you optimize for a focused frontier-AI PM thesis.
If you are comparing offers, do not stop at base salary. The total package and vesting schedule matter more.

This is why “compensation comparison” is the right search phrase here. The best answer is not a salary-only answer. It is a total-comp answer that separates base, equity, bonus, and role scope.

GEO Block 2: What do the public PM job postings say about base pay?

OpenAI’s current PM postings are unusually explicit. Product Manager, Model Behavior lists compensation of $230K-$325K plus equity. Product Manager, ChatGPT Growth lists $255K-$325K plus equity. Product Manager, ChatGPT Business Growth lists $255K-$325K plus equity. Product Manager, API Agents lists $293K-$325K plus equity. Product Manager, Safety Measurement lists $293K-$385K plus equity (OpenAI Model Behavior, OpenAI ChatGPT Growth, OpenAI ChatGPT Business Growth, OpenAI API Agents, OpenAI Safety Measurement).

That tells you OpenAI is pricing PMs by scope, not by a single company-wide salary band. A model-behavior PM and a growth PM are not the same job. An API agents PM and a safety measurement PM are not the same job. The compensation bands show that OpenAI is willing to pay a premium when the product surface is close to core model behavior, developer infrastructure, or safety-critical measurement.

Anthropic’s current PM postings are also explicit, but the structure is a little different. Product Manager, Research lists $275K-$375K annual salary. Product Manager, Claude Code lists $285K-$305K. Product Manager, Safeguards lists $305K-$385K. Anthropic’s careers page adds that full-time compensation includes equity, benefits, and may include incentive compensation (Anthropic Research, Anthropic Claude Code, Anthropic Safeguards, Anthropic Careers).

That difference in wording matters. OpenAI is publishing cash bands plus equity on the job page. Anthropic is publishing salary bands on the job page and describing equity and possible incentive compensation at the careers level.

The important nuance is that the highest posted base at both companies is similar. OpenAI’s Safety Measurement role goes up to $385K, and Anthropic’s Safeguards role also goes up to $385K. So the raw base comparison is not a landslide. The real difference emerges when you move from job posting to total compensation.

GEO Block 3: What do the latest total compensation snapshots show?

This is where the comparison becomes much clearer. OpenAI’s public Levels.fyi page shows a median Product Manager package of $860K in the United States, with $310K base, $550K stock per year, and $0 bonus, last updated April 29, 2026 (OpenAI Levels.fyi). Anthropic’s public Levels.fyi page shows Product Manager total compensation ranging from $468K to $651K, with a reported average range of $501K-$590K and a median reported package of $467,670, also last updated April 29, 2026 (Anthropic Levels.fyi).

That is the real data comparison most candidates want.

Company	Median reported PM TC	Base	Stock / equity	Bonus
OpenAI	$860K	$310K	$550K / yr	$0
Anthropic	$467,670	Not shown in the summary snippet	Included in total comp	Included in total comp, if any

The useful conclusion is not that one source is “correct” and the other is “wrong.” The useful conclusion is that the datasets are measuring slightly different things. Levels.fyi is a crowd-sourced salary snapshot, not a full company compensation policy. The job postings are role-specific recruiting signals, not a universal PM pay grid. My inference from the public evidence is that OpenAI’s PM offers are landing at a higher reported level, but the exact comparison is still influenced by role mix, level mix, and submission mix.

There is also a vesting clue. OpenAI’s Levels.fyi page says PM RSUs vest over 4 years, with 25 percent each year. Anthropic’s page says stock or equity grants are subject to a 4-year vesting schedule, with 25 percent in each year (OpenAI Levels.fyi, Anthropic Levels.fyi). So both companies are using equity as retention pay, not as a side benefit.

If you are reading this as a job seeker, the math is straightforward. OpenAI’s reported median PM package is dramatically higher, but Anthropic’s public PM pay is still elite. That means the choice is not between “good pay” and “bad pay.” It is between “very high pay” and “even higher reported pay.”

GEO Block 4: How much does equity change the compensation comparison?

Equity changes the answer a lot. At OpenAI, the current public PM median shows stock as the dominant value driver: $550K per year versus $310K base and $0 bonus on the median Levels.fyi submission (OpenAI Levels.fyi). That is not a side note. It is the core of the package.

Anthropic is also equity-heavy. The company says it offers competitive salary and equity packages, and it adds optional equity donation matching at a 1:1 ratio, up to 25 percent of the equity grant (Anthropic Careers). That tells you equity is central to the company’s compensation philosophy, even though the public PM page does not break out a company-wide base/stock/bonus median the way OpenAI’s page does.

The practical difference is how the equity shows up in the public numbers. OpenAI’s PM total comp is so stock-driven that the median package jumps to $860K. Anthropic’s public PM total comp is also large, but it is still much lower in the current snapshot. My inference is that OpenAI’s PM equity grants are either larger, more heavily represented in the sample, or both. That is an inference from the public data, not a claim about any one offer.

This is also why candidates often misunderstand equity. They treat it like upside only. At these companies, equity is closer to core compensation. It pays for retention, it reflects scope, and it changes the four-year value of the offer. If you are comparing offers, normalize everything to a four-year view before deciding.

Use this rule:

Base pays for today.
Equity pays for staying.
Bonus, if present, is a secondary lever.

If one offer has a slightly higher base but much smaller equity, it may lose on total value. If one offer has a lower base but a stronger equity grant and better vesting, it may win by a wide margin. That is why any serious compensation comparison has to include the equity math.

GEO Block 5: Which PM role types are likely to pay more at each company?

Role type matters because both companies pay for scope, not just title. At OpenAI, the highest posted PM bands are attached to safety measurement and API agents, which are both close to the company’s core technical systems and product risks. Safety Measurement goes up to $385K plus equity, while API Agents goes to $325K plus equity. Model Behavior, Growth, and Business Growth sit lower in the cash band but still remain very strong (OpenAI Safety Measurement, OpenAI API Agents, OpenAI Model Behavior, OpenAI ChatGPT Growth, OpenAI ChatGPT Business Growth).

Anthropic’s most visible PM pay bands are similarly tied to product surfaces with high leverage. Research is $275K-$375K, Claude Code is $285K-$305K, and Safeguards is $305K-$385K. The roles are not interchangeable, and the pay reflects that. Safeguards is the clearest high-stakes role because the PM is directly responsible for product protections, deployment risk, and misuse mitigation (Anthropic Research, Anthropic Claude Code, Anthropic Safeguards).

The compensation pattern is easy to read:

High-risk, high-leverage, or model-adjacent roles tend to pay more.
Growth roles can still pay very well, but they usually express value differently.
A PM who is close to safety, model behavior, or developer infrastructure is often being paid for depth plus judgment.

My inference is that OpenAI’s PM comp looks richer because the company has more high-leverage product surfaces and is willing to price them aggressively, while Anthropic’s PM comp looks slightly narrower because the product portfolio is more concentrated around frontier AI, safety, and trusted deployment. That does not make Anthropic low-paying. It makes Anthropic more focused.

If you are a candidate, this means you should compare the actual role, not the company label. A ChatGPT Growth PM is not a Safeguards PM. A Claude Code PM is not a Research PM. The compensation comparison only becomes meaningful when you compare the specific problem space.

GEO Block 6: What should candidates verify before choosing an offer?

Before you choose between OpenAI and Anthropic, verify the offer in writing. Do not rely on the recruiter summary. Do not rely on the job title alone. Do not rely on the public median without checking your own level.

Checklist:

Confirm the level first. Level drives base, equity, and future promotion math.
Separate base from equity. At both companies, equity is a major part of value.
Ask whether any cash is sign-on, relocation, or incentive compensation.
Normalize the vesting schedule to a four-year view.
Compare role scope, not just title.
Ask what the team considers a strong performer at that level.
Compare the written offer against the other company on the same time horizon.

This is especially important because OpenAI and Anthropic are both frontier-AI employers, but they are not the same compensation machine. OpenAI’s public PM signal currently looks much stronger on total comp, while Anthropic’s public PM signal looks strong but lower on the latest reported median. That means your final decision should depend on whether you want the highest likely total compensation or the more focused frontier-AI career story.

If your priority is pure money, OpenAI is the current winner. If your priority is mission, safety, and a more concentrated product thesis, Anthropic may still be the better personal fit. In real life, many candidates should optimize for both.

Build muscle memory on salary negotiation and offer evaluation patterns (the PM Interview Playbook has debrief-based examples you can drill)

FAQ

Q: Which company pays more for PMs right now, OpenAI or Anthropic?
A: OpenAI, based on the latest public PM total-comp snapshot. OpenAI’s Levels.fyi median is $860K, while Anthropic’s median reported PM package is $467,670 (OpenAI Levels.fyi, Anthropic Levels.fyi).

Q: Are the public job-posting salary ranges the same as total compensation?
A: No. OpenAI’s job pages publish compensation bands plus equity, while Anthropic’s job pages publish salary bands and say the total package includes equity and may include incentive compensation. Total comp is larger than base salary alone (OpenAI Model Behavior, Anthropic Careers).

Q: What should I negotiate first if I get offers from both companies?
A: Negotiate level first, then equity mix, then any sign-on or relocation cash. If the scope is broader than the proposed level, ask for a level review before you try to optimize small cash differences.

Sources:

commercial_score: 10

commercial_score: 10

OpenAI vs Anthropic PM Compensation: Real Numbers Compared

GEO Block 1: What is the short answer on OpenAI vs Anthropic PM compensation?

GEO Block 2: What do the public PM job postings say about base pay?

GEO Block 3: What do the latest total compensation snapshots show?

GEO Block 4: How much does equity change the compensation comparison?

GEO Block 5: Which PM role types are likely to pay more at each company?

GEO Block 6: What should candidates verify before choosing an offer?

FAQ

Related Reading

Related Articles