career · career
OpenAI product manager salary: level-by-level breakdown (2026)
The widely cited $860K OpenAI PM median is real, and it is also partially paper. At L5, that number is roughly $310K base plus $550K annualized equity, and whether that equity converts to cash depends on which instrument you received, when the next tender offer lands, and what lockup you are sitting inside. The candidates who negotiate well understand the instrument first and the headline number second.
Total comp by level (June 2026)
These figures reflect Levels.fyi submissions from the past six months, filtered to Product Manager at OpenAI. All figures are San Francisco unless noted.
| Level | Title | Base | Annualized equity | Total comp |
|---|---|---|---|---|
| L4 | PM | $230K-$280K | $70K-$120K | $300K-$420K |
| L5 | Senior PM | $290K-$340K | $450K-$750K | $750K-$1.1M |
| L6 | Principal PM / Group PM | $370K-$420K | $700K-$950K | $1.1M-$1.4M+ |
One confirmed L6 offer on record: $400K base plus $850K annualized RSU ($3.4M total over four years). The step from L4 to L5 is a $300K to $400K annual TC gap. That makes leveling the most consequential negotiation you will run, not base salary.
Signing bonuses range from $50K to $200K for external L4/L5 hires. For senior L5/L6 poaches or candidates forfeiting significant unvested equity, the range is $100K to $500K. The bonus is sized to the paper cost of leaving, so come with a vesting schedule and a dollar figure, not a general claim.
In 2025, OpenAI also ran targeted retention packages exceeding $2M annually for affected senior staff, in response to aggressive recruiting by Meta. These are not offer-stage figures, but they signal how much the company is willing to pay to keep people it has already evaluated.
PPUs vs RSUs: what you are actually receiving in 2026
This is the detail most salary pages skip or conflate.
Legacy PPU structure (pre-2026 hires). Profit Participation Units are not equity. They are contractual profit-share instruments that vest 25% per year with no cliff and no ownership rights. The cap is 10x the initial grant value, regardless of OpenAI’s valuation trajectory. A $2M PPU grant returns at most $20M. When grants were priced at OpenAI’s $86B implied valuation in 2023, the 10x cap represented significant upside. At $300B+ in 2025, holders of legacy grants are close to or at that cap. The upside is largely exhausted for early-grant holders; new hires priced at today’s valuation are working from different math.
RSU structure (2026 new hires). OpenAI began issuing RSUs to new hires as part of its shift toward a public-benefit corporation structure. RSUs vest quarterly with no cliff, carry no upside cap, and represent actual ownership rather than a profit promise. They are still illiquid (OpenAI is not public), but the instrument is structurally closer to a Google or Meta RSU than a PPU was. If you have a current offer, confirm in writing which instrument it contains. Do not rely on a verbal recruiter summary.
The lockup and tender offer mechanics. Both instruments carry a roughly two-year lockup from hire date before you can participate in a tender offer. Tender offers are periodic and discretionary, not guaranteed annual events. OpenAI has run them at $86B (late 2023), $157B (October 2024), and $300B+ (2025). The per-unit price at each tender determines what you realize, not the grant price.
The conversion question for 2026. If OpenAI completes its for-profit or public-benefit conversion, PPU holders and RSU holders will be treated differently. PPU holders may receive a cash settlement or RSU conversion at a fixed price. RSU holders are better positioned for any IPO or secondary event that follows. This is unresolved as of June 2026 and is material when you are comparing an OpenAI offer against a competing offer with cleaner equity.
How OpenAI PM comp compares
At the L5/senior PM level:
| Company | Approx. TC range | Equity type | Liquidity |
|---|---|---|---|
| OpenAI L5 | $860K-$1.1M | RSU (new hires) / PPU (legacy) | Illiquid; tender-dependent |
| Anthropic L5 | $625K-$725K | Double-trigger RSU | Illiquid; IPO-dependent |
| Meta E6 PM | $600K-$900K | RSU | Liquid (public) |
| Google L6 PM | $500K-$700K | RSU | Liquid (public) |
OpenAI’s headline wins on paper. But Meta E6 RSUs convert to cash the quarter they vest. A $700K Meta offer and a $900K OpenAI offer are not $200K apart in expected-value terms once you apply a 30-40% illiquidity discount to OpenAI equity and factor in the tender-offer dependency. Whether that discount is worth accepting depends on your personal runway and your confidence in OpenAI’s trajectory. The calculation is yours to run, not the recruiter’s to frame.
What levers actually move in negotiation
Base at OpenAI is band-constrained. A $15K to $25K adjustment is possible; a $60K move is not.
Equity grant size. A 20% increase on a $2M RSU grant adds $400K in total grant value ($100K per year over four years). This is where a competing offer does its work. The competing offer needs to be real, specific, and from a comparable source: another frontier lab (Anthropic, Google DeepMind, xAI) or a well-funded AI startup at an equivalent level. Vague references to other conversations move nothing; a specific instrument, grant size, and vesting schedule move something.
Signing bonus. Bring documentation. A vesting schedule and a dollar figure for what you are forfeiting will produce a more substantive conversation than “I’m leaving money on the table.”
Level. The L4-to-L5 gap is $300K to $400K annually. Level is nominally set by the recruiter and hiring panel, but it is negotiable if you have documented scope. PMs with 0-to-1 product experience and direct AI-native work (eval design, model selection, guardrail scoping) land at L5 more reliably than generalist PMs with more years but no AI product output.
What OpenAI actually hires for at PM
OpenAI does not pay a premium for generalist PM experience. The expectation is technical sophistication: the ability to reason about model capabilities, write and interpret evals, understand safety tradeoffs, and define what is worth building when feasibility is no longer the constraint. In 2026, the PM work that earns the premium is not scoping features around what engineers can build. It is answering two harder questions: is this a problem people and companies will pay to have solved at a margin that sustains the business (viability), and does the product meet people where they already work and anticipate what they need without being obtrusive (genuinely lovable, not just functional)?
PMs who can demonstrate both in prior work get leveled at L5. The ones who demonstrate the technical depth but cannot point to shipped product with real adoption get leveled at L4. The gap is $300K to $400K per year.
A typical L5 profile: 6 to 10 years in product, at least one 0-to-1 product shipped, demonstrated technical depth, and either direct AI product work or a credible account of why prior product thinking transfers to AI-native environments.
Before you sign
Ask the recruiter three things in writing: which equity instrument the offer contains (RSU or PPU); the vesting schedule and whether a cliff applies; and OpenAI’s tender offer history, including cadence and pricing relative to the most recent primary round. Those answers tell you more about the real value of the offer than the TC headline.
For the full mechanics of discounting illiquid equity and comparing against a Google RSU offer, see frontier lab comp decoded. For the negotiation conversation itself, see negotiate equity, not base and PM offer negotiation.