Orixa Honeypot Report: What 41,536 Scans Show

A honeypot lets you buy a token but blocks — or economically destroys — your ability to sell it. This report summarizes what Orixa's own scan history actually shows about how often that pattern appears, on which chains, and in what form. Every figure below is drawn from stored scan decisions, not estimates. Read them as flags, not verdicts: a honeypot flag reflects the on-chain and simulation evidence available at scan time, it is not by itself proof of intent to defraud, and automated detection can produce false positives.

The dataset

As of August 2026, Orixa's scan ledger holds decisions for 41,536 distinct tokens across BNB Chain, Base, Ethereum, and several smaller networks. Of those, 5,707 were classified very-high-risk (a risk score of 75 or above). The honeypot findings in this report are measured against that population — an append-only record of what each scan saw at the moment it ran.

Coverage is uneven by design: it reflects which tokens were submitted for analysis, not a random sample of every deployment. So these numbers describe the tokens Orixa was asked about, which skews toward new, low-liquidity launches — exactly where honeypots concentrate.

How often the honeypot pattern appears

Two independent signals mark a token as effectively unsellable, and this report separates them because they mean different things.

Sell-simulation failure. Orixa runs an on-chain sell simulation on supported networks. Among the very-high-risk population, 153 tokens carried a honeypot flag from a failed or blocked sell — the analysis could complete a buy but not a sale. By chain, those split 76 on Ethereum, 62 on BNB Chain, and 15 on Base.

Punitive sell tax. A sale can technically succeed while the contract keeps almost everything. Of the tokens where a sell tax could be measured, 26 charged 50% or more, and 21 charged a full 100% — a sale returns nothing. Economically, a 100% sell tax is a honeypot even when the transaction does not revert. (The full spread is covered in Crypto Sell-Tax Distribution: What We Found.)

These two groups overlap only partly: a token can fail a sell simulation with a 0% listed tax, or pass the simulation while charging a confiscatory tax. Counting either signal alone understates the problem, which is why a scanner should test the sell path directly rather than trust a single field.

Why the raw "is honeypot" boolean isn't enough

A single provider boolean labelled honeypot is unreliable on its own — across the same ledger it was rarely set, even for tokens that were mechanically unsellable. The decisive evidence came from Orixa's own sell simulation and the measured tax, not a borrowed flag. This is the core lesson of the data: sellability has to be tested, per token, per pool, at scan time — a label copied from elsewhere misses most of the cases and mislabels some safe ones.

What this does and does not claim

• It does not claim these 153 tokens are proven frauds. A flag is a reason to stop and verify, not a conviction. Some may be misconfigured, paused pre-launch, or false positives.
• It does not claim a honeypot rate for crypto as a whole. The sample is tokens submitted to Orixa, weighted toward risky launches.
• It does show that among tokens already flagged high-risk, unsellability is common and takes more than one form — and that testing the sell path is what surfaces it.

See the live evidence

The named, current entries behind these counts are published — with the same "flags, not verdicts" framing — in Orixa's public Tokens Flagged as Very High Risk report, generated only from stored scan decisions. To understand the mechanism, see How to Check If a Token Is a Honeypot; to read a single detection end to end, see A Honeypot Case Study: Reading the Evidence.

Figures are from Orixa's scan ledger as of August 2026 and change as new tokens are scanned and re-scanned. Run a contract address through Orixa to test its sell path yourself, then verify anything critical on a block explorer. Orixa is decision support, not a guarantee of safety.