At E5, Meta paid a median of $501,000 last year. Google L5, the same level, paid $409,000. That's a $92,000 gap for the same job title, before you account for vesting schedules or refresh cycles. The "FAANG pays well" line collapses pretty quickly once you compare same-level offers side by side.

All numbers below come from the Levels.fyi 2025 End-of-Year Pay Report and the per-company salary pages on Levels.fyi. Total comp here means base + annualised stock + target bonus, the way Levels.fyi reports it. If you compare against an offer letter using "base + RSU at grant", you'll undercount.

Median total comp by company, 2026

Meta sits at the top of the median table for one specific reason. After the META rebrand and the 2024 stock recovery, RSU grants annualise much higher than they did during the 2022 dip, and refreshers are coming in at levels Meta hadn't paid since pre-pandemic. Whether the stock holds is an open question. (At E6 the picture is even more lopsided. Meta's median is $788k, Google's is $537k.)

Level equivalents matter more than the company name

Most candidates compare offers wrong. They look at the all-levels median and pick the highest. The right comparison is same-level to same-level, because a "Google L4" is a "Meta E4" is an "Amazon L5", and conflating titles loses you tens of thousands of dollars a year. Here's the rough map.

Two things to flag. First, Amazon's level numbers don't map one-to-one onto seniority the way Google's do (Amazon's L4 is entry, Google's L4 is mid). Second, Netflix doesn't level publicly. They hire "senior" and above, with title compression that doesn't translate cleanly to any other company.

The vesting schedule is half the story

Two engineers with the same headline TC can take home very different cash over four years depending on how the stock vests. The schedules vary more than people realise.

Amazon's schedule is the trap candidates get caught in most often. The offer letter quotes total comp annualised across four years, but you only get 20% of the stock in the first two years. Signing bonuses bridge the gap. If you leave before year three, you've earned closer to entry-level than the headline number. There's a reason Amazon's median tenure for SWEs is shorter than the rest of FAANG.

Google by level

Google's ladder runs L3 (entry) through L11 (Fellow). Most engineers cap out at L5 or L6. L4 is now a terminal level, meaning staying there indefinitely doesn't push you out the door.

Meta by level

Meta pays roughly 15 to 25% more than Google at every equivalent level. The trade-off is the review culture. Meta's twice-yearly PSC calibration is more rigorous than Google's, and a "Meets Expectations" rating sets off a real conversation with your manager. Engineers who came over from Google in 2024 and 2025 talk about this constantly.

E6 to E7 is the steepest jump in tech. Most of that gain is stock (annualised RSU grants north of $1.2M at the top of band). Fewer than 1% of engineers reach E7. If you're being recruited at that level you already know the comp will be custom.

Amazon by level (with the back-loaded asterisk)

Apple and Netflix: the outliers

Apple pays less than Meta or Google at every level, by 10 to 15% on average. ICT4 medians around $380k. The pitch is brand, stability, and the smallest gap between IC pay and manager pay in FAANG. Apple also famously runs a tighter return-to-office policy than its peers, which matters more to some candidates than the comp gap.

Netflix is its own model. No RSUs, no bonus, no four-year vesting at all. They pay one number, in cash, every payroll cycle, and you can elect to convert part of it into stock purchases if you want. Senior SWEs range from $350k to $500k all-cash. Staff is $500k to $700k. Principal is $700k+. The upside is predictability. The downside is that you don't participate in upside if Netflix stock doubles. The other downside is the "keeper test": Netflix expects managers to continuously ask whether they'd fight to keep each report, and "adequate performance gets a generous severance" is part of the actual handbook.

What we hear from candidates comparing offers

Candidates working with the LastRound AI copilot on offer prep get the same question wrong over and over: they optimise for headline TC instead of effective TC. Headline TC is what the recruiter quotes. Effective TC is what you actually take home over the years you realistically stay, accounting for vesting schedule, refresh patterns, base-pay caps, and the odds you'll survive the performance-review cycle.

A few rules that hold up in almost every comparison we walk through:

• If you're optimising for cash in years 1 and 2, Google wins. The front-loaded vesting plus a reasonable base does more than Meta's higher headline.
• If you're optimising for four-year max TC and you trust the stock, Meta wins.
• If you're optimising for predictability or you can't stomach RSU volatility, Netflix wins.
• If you're optimising for brand or work-life balance, Apple or Google wins (team-dependent).
• Amazon wins on offer-letter total comp and signing bonuses, but rarely on take-home cash if you exit before year 3.

How to negotiate between FAANG offers

Five things that actually move the needle in practice:

1. Get the level right first. An L5 at one company can outpay an L6 at another. Recruiters can match or beat a competing level, which is a bigger lever than a competing dollar figure.
2. Compare four-year totals, not year-one. Amazon's year-one number looks competitive only because the signing bonus is doing the work.
3. Use competing offers literally. Recruiters at the FAANGs all have explicit latitude to match. They need you to forward the other offer letter or the recruiter's number.
4. Push on signing bonus and refresher grant. These are the most flexible levers. Base salary is band-locked, initial stock is band-locked, but signing and refreshers are negotiable.
5. Factor in the keep-or-lose probability. A higher headline TC at a company with a 25% PIP rate is worth less than a 15% lower TC somewhere stable. (This is the one nobody on Reddit will admit to you.)

One more thing

All of the above is the candidate side. From the recruiter side, the conversation looks different in 2026. Tech hiring budgets at FAANG are not back to their 2022 peaks. The BLS Software Developers Occupational Outlook still projects strong growth, but the FAANGs themselves have fewer open requisitions per quarter than they did in 2021. That changes the negotiating dynamic. If you have a strong offer in hand, negotiate hard. If you don't, applying broadly matters more than chasing the highest single offer.