Introduction
Prompt engineering pay turned into dinner‑table trivia in 2023—fueled by splashy listings and a new class of AI roles. If you’re job hunting (or budgeting headcount), you’ve probably landed on Levels.fyi to decode what “prompt engineer” actually earns in 2025. We tested the platform specifically for prompt/LLM-focused roles and compared it with Glassdoor, Payscale, Indeed, and a sampling of public postings to see whether Levels.fyi gives a trustworthy answer—and how to use it without getting misled.
What We Reviewed (and How)
Levels.fyi is a crowdsourced compensation platform popular with engineers and product folks. You can filter by title, level, company, location, and sometimes by specialization (e.g., AI/ML). For prompt engineering specifically, job titles vary: you’ll see “Prompt Engineer,” “LLM Engineer,” “AI Applications Engineer,” and “GenAI Engineer.” We used those variants when assessing data quality.
Where possible, we cross-checked against:
- Glassdoor’s salary pages for prompt engineers and adjacent roles (Glassdoor).
- Hired’s State of Software Engineers (AI/ML sections) for directional pay ranges (Hired).
- Payscale and Indeed for broader market medians (Payscale, Indeed).
- Public postings that disclosed ranges (e.g., Anthropic’s notable “Prompt Engineer & Librarian” role reported at $175k–$335k base in 2023; The Verge coverage).
US prompt engineer pay snapshot (triangulated)
| Segment | Typical Base Range | Typical Total Comp | Notes |
|---|---|---|---|
| Mid-market/remote AI app teams | $120k–$170k | $140k–$220k | Broader title variance; equity small-to-moderate |
| Big Tech/FAANG AI app roles | $160k–$230k | $200k–$350k | Equity and bonuses drive variance |
| Top-tier AI labs/startups | $200k–$300k | $250k–$500k+ | Outliers exist; scope/impact heavy |
These ranges synthesize self-reported data on Levels.fyi with directional ranges visible on Glassdoor/Indeed and public job posts as of late 2024. Your mileage will vary by title, scope, and location.
What Levels.fyi Gets Right (Pros)
- Real-time, crowd-powered signal: New AI roles change fast. Levels.fyi’s velocity of submissions makes it one of the earliest places you’ll see pay shifts for specialized AI work.
- Equity clarity: Where available, the breakdown of base, equity, and bonus helps you compare a $190k base + low equity offer vs. a $165k base + high upside package.
- Granular filters: Sorting by location (SF Bay, NYC, remote), company, and level is crucial in AI where pay swings are large.
- Community norm-setting: Seeing bands and datapoints reduces information asymmetry during negotiation.
TipHow to get cleaner comps
- Filter by location and company size.
- Include adjacent titles (LLM Engineer, AI Applications Engineer).
- Discard entries without breakdowns or that look like clear outliers.
- Compare at least 3 sources before anchoring.
Where It Falls Short (Cons)
- Sample bias at the top end: AI-curious candidates at well-known companies are more likely to report, skewing ranges higher than the broader market.
- Title ambiguity: Self-entered titles (“Prompt Engineer,” “AI Engineer,” “ML Engineer”) can blur categories and inflate/deflate ranges if you don’t filter carefully.
- Inconsistent verification: While Levels.fyi does moderation and some checks, self-reported data can include stale, mis-leveled, or unverified entries.
- Thin coverage outside US tech hubs: International and non–Big Tech data can be sparse, which matters if you’re outside SF/NYC or in hybrid roles.
Verdict and How to Use It
If you’re asking “What’s a prompt engineer salary?” Levels.fyi is one of the best starting points—especially when paired with Glassdoor, Indeed, and current postings. We’d call it a 4/5 for this use case: excellent directional truth, but not a single source of truth.
Actionable steps to price your market value:
- Define your lane: Are you doing pure prompt design, or full-stack LLM app work (RAG, tools, evals, observability)? Full-stack usually pays more.
- Triangulate: Pull a range for your location and level on Levels.fyi, then cross-check with Glassdoor and at least two live postings with disclosed ranges.
- Normalize by company stage: Early-stage startups may trade cash for equity; public companies lean heavier on RSUs and bonuses.
- Negotiate with components: If base is capped, push on equity refreshers, signing bonus, or role scope/title to match market peers.
Bottom line: For US roles in 2025, prompt/LLM-focused engineers commonly see base salaries in the $140k–$220k range, with total compensation stretching to $300k+ at brand-name tech and frontier labs. Exceptional packages exist, but they’re tied to clear impact and specialized skills. Use Levels.fyi to set your anchor—then validate twice before you sign.
