What does LLM inference cost an agent per call over x402?
The median LLM inference provider on x402 charges $0.02 per call. The middle half of the market sits between $0.005 and $0.1, based on 933 distinct priced offers from 152 providers. Updated 2026-08-05.
Chat completions, embeddings and model routing billed per request in USDC instead of a prepaid credit balance.
Where a price would land
Each rung is the provider-weighted percentile across this category. Pick the position you want to hold, then read the price.
Price distribution
Distinct priced offers per bracket — templated route families collapsed so one product doesn’t count hundreds of times.
⚠ concentration: llm402.ai publishes 40% of the distinct offers in this category. The provider-weighted median above corrects for this; the raw per-route median is $0.008.
Provider comps
What each seller actually publishes in this category, ranked by catalogue size.
Live routes across the price range
Real endpoints sampled from the cheapest to the most expensive offer in this category.
Methodology
Every figure is built from prices providers publish for their own endpoints, captured by the TOLL·402 discovery crawl and rebuilt 2026-08-05. Where a strict HTTP 402 check succeeded we use the price the endpoint actually quoted rather than its catalogue entry — 401 of the 933 offers here are priced from a live quote. That matters: across the corpus, roughly one in twenty-seven catalogued prices disagrees with the live quote, some by more than 100×.
Two corrections are then applied before any median is taken. First, templated route families — one endpoint per card, per ticker, per city — are collapsed into distinct offers keyed by provider, title and price, so a 900-route family counts as one price point. Second, each provider contributes the median of its own offers exactly once, so a single bulk seller cannot move the benchmark.
Prices are what providers publish in their 402 quote. They are not evidence of demand, and a confirmed quote is not a settled payment. A category is only published once at least 4 distinct providers are observed in it.
FAQ
How much should I charge for a LLM inference API on x402?
Across 152 providers publishing LLM inference routes, the median seller charges $0.02 per call. The middle half of the market sits between $0.005 and $0.1. Pricing at or just below the median is the least friction for a new listing; above $0.5 you are at the top of the observed market and need a clear capability reason.
Where do these numbers come from?
TOLL·402 crawls the public x402 discovery surface and records the price each endpoint publishes in its HTTP 402 quote. This page covers 933 distinct priced offers from 152 providers, last rebuilt 2026-08-05. Prices are what sellers publish — not settled payments and not evidence of demand.
Why is the median different from the average route price?
Because the corpus is concentrated. A single provider can register thousands of near-identical routes and drag any per-route average toward its own list price. We collapse templated route families into distinct offers, then let each provider contribute its own median exactly once — so this benchmark reflects what a typical seller charges, not what the largest seller charges.