How to read a Solana token's risk score (grades A–F)

November 10, 2026 · 12 min read · by the Rugsight team

A rug-check score is easy to misuse. People see a letter, treat A as a green light or D as proof of a scam, and skip the flags that produced the number. On Rugsight the score is narrower than that: it is an automated 0–100 technical summary of public on-chain and market checks, then a letter grade. This guide explains how to read that grade, what each band means, and what it does not certify, with live examples checked on 10 November 2026 and re-checked the same evening (23:18–23:38 America/Fortaleza) where a figure had moved.

The checks behind the number are walked through in how to check a Solana token for rug-pull risk. Issuer-controlled assets (stablecoins, some wrapped tokens) use a separate label, explained in what issuer-controlled means.

What the score measures

Rugsight starts at 100 and subtracts points for the flags it can confirm, then adds a small bonus only when it can verify that a pool's LP is burned or durably locked. The usual deductions cover:

• Authorities: mint authority still active (30), freeze authority still active (20).
• Token-2022 controls such as a permanent delegate, transfer hook, pause authority, default-frozen accounts, or a transfer fee. Details are in how to check Token-2022 permanent delegate and hooks.
• Holder concentration among real wallets, not raw token accounts: for example TOP10_ELEVATED (top 10 over 30%) is 10 points, TOP10_HIGH (over 50%) is 20, TOP10_EXTREME (over 80%) is 30; a single wallet over 10% (TOP1_LARGE) is 8, over 20% (TOP1_WHALE) is 15. See how to check holder concentration.
• Liquidity and pool age: no pool (25), very low liquidity (20), low (10), still on a pump.fun bonding curve (10), or a modest-liquidity flag (5); trading age costs 15 under 24 hours, 8 under 7 days, 3 under 30 days.

If holder data or market data cannot be read, the score is capped at 84 (grade B). The tool will not call a token "low risk" on the checks it runs when a major input is missing.

The number is a snapshot of those checks at checkedAt. It is not a price forecast, a legitimacy badge, or a statement about the team.

The grade bands

The public letter is derived from the numeric score, except when the mint is on the issuer-controlled allowlist:

An A at exactly 85 is still an A. A hours-old token with only the VERY_NEW flag (15 points) can land on that line if nothing else fires — which is why you should read the flags, not only the letter. See how to check token age.

IC replaces the letter when the mint and its live authorities match a short allowlist (USDC, USDT, some wrapped assets, tokenized stocks). The technical score is still calculated and still deducts for those controls. If an authority moves to an address we did not pin, the IC label is dropped and a normal letter is used.

How to read a score yourself

1. Confirm you have the right mint, not a same-ticker lookalike (verify the contract address).
2. Paste it into Rugsight or open the token page if it is already in the popular list.
3. Read the grade and score first, then the verdict, then every flag. The letter is a summary; the flags are the reason.
4. If the grade is IC, read the issuer note and the technical score. Do not treat IC as an A.
5. If two mints are in the same category, put them side by side on a comparison page (for example BONK vs WIF or USDC vs PYUSD).
6. Use the full checklist for anything the score does not cover (deployer history, socials, intent).

Live examples (verified 2026-11-10, re-checked that evening)

These numbers come from Rugsight's own free score on 10 November 2026. Major-token rows were re-checked at 23:18–23:38 America/Fortaleza (11 November 2026, 02:18–02:38 UTC). They will change; open the linked token page for the live figure.

Grade A — few or no deductions:

TRUMP is a clean 100 on the checks we run: mint and freeze are revoked and no warning flags fired. POPCAT is an A because a single low-severity holder flag (TOP10_ELEVATED, 10 points) is not enough to leave the A band. WIF sits on the A/B line (85) with the same two holder flags that put BONK and Pyth at B / 82 in the next table; WIF's main Raydium AMM v4 pool also showed 99.75% of LP burned, which is a small verified bonus, while Pyth's main pool is a concentrated Raydium CLMM where an LP-burn percentage does not apply. An A is "the technical traps we score were mostly absent", not a recommendation to buy.

Grade B and C — well-known tokens, not new memecoins:

BONK was A / 90 earlier on 10 November 2026, with only TOP10_ELEVATED. By 23:18 America/Fortaleza the same day the largest real wallet had crossed 10% (TOP1_LARGE) and the score was B / 82 — the same two holder flags as Pyth. That is why you should read the flags and the timestamp, not only the letter. JUP, ORCA and RAY have been trading for a long time and are widely used. They are still C-grade on this check because of holder concentration (JUP, RAY) or a still-active mint authority plus concentration (ORCA). A C on a familiar ticker means several of the technical checks fired, not that the project has been shown to be a scam, and not that the largest wallets are necessarily a hidden deployer. Compare two of them at JUP vs ORCA.

Grade D — serious technical flags, including on a known mint:

• RENDER (rndrizKT3MK1iimdxRdWabcF7Zg7AR5T4nud4EkHBof): grade D, score 32 (free check). Flags: MINT_AUTHORITY_ACTIVE, FREEZE_AUTHORITY_ACTIVE, TOP10_ELEVATED, TOP1_LARGE. Those authority flags are the same shape you see on USDC, but RENDER is not on the issuer-controlled allowlist, so they count as a letter grade instead of IC. A D here is a statement about the controls we measure, not a claim that the project is fraudulent.
• PVE (5rT4jYmYMZEkkDgiMwE3dRvL3vgy7dZovSi4RaJmL1jC): grade D, score 30. Flags: TOP10_EXTREME, TOP1_WHALE, VERY_NEW, LIQUIDITY_LOW. Earlier the same day, in the age guide, this mint scored F / 20 with LIQUIDITY_VERY_LOW instead of LIQUIDITY_LOW. The score moved when the inputs moved. Treat every grade as a timestamped snapshot.

Grade F — several high-severity flags at once:

• FomoInu (8i32WG6bMAzztdZB4NLbFw4sVo2d4VKCRyuL53A3Bce5): grade F, score 20 (free check). Flags: TOP10_EXTREME, LIQUIDITY_VERY_LOW, TOP1_WHALE, VERY_NEW. A single large wallet, thin liquidity and a pool less than a day old is a typical F-shaped new launch on the new-token feed.
• JENSEN (5kop9u6Dg5iSmJ3qBnL3SsazBZkcAY3nw3wJrAq4h2gK): grade F, score 28. Flags: MINT_AUTHORITY_ACTIVE, VERY_NEW, TOP10_ELEVATED, TOP1_LARGE, LIQUIDITY_MODERATE, TRANSFER_FEE. This F is a different mix: the mint can still print supply, the pool is new, and a Token-2022 transfer fee is on. Liquidity here was only flagged as moderate, so the failing grade is not "no pool" — it is the combination.

Grade IC — issuer-controlled, technical score still visible:

USDC's technical 50 is almost entirely the mint and freeze deductions (30 + 20). Holders were not checked on that run, so they did not add or subtract. cbBTC's technical 15 adds concentrated holders on top of the same authority flags. Both stay IC because the issuer addresses match the allowlist; the low technical score is there so the controls stay visible. Side by side: USDC vs PYUSD.

What Rugsight shows

On the free check:

• the grade (A–F, or IC) and the 0–100 score;
• a one-line verdict that matches the bands above;
• the individual flags, which you should read before the letter;
• for IC mints, the issuer name, asset type and the technical score that would have been the letter grade.

The paid report (1 USDC for people, or 0.03 USDC per call for agents via x402) adds the exact figures behind holder and market buckets and, on young tokens, a launch-bundle / sniper heuristic. The free BONK sample shows the layout.

Limits, stated plainly:

• The score only covers the checks listed above. It does not score deployer history, socials, whether a large wallet is a foundation treasury, or whether early buyers will sell.
• Missing holder or market data caps the score at 84. A "B" in that situation can mean "we could not finish the check".
• LP burn or lock only changes the score when Rugsight can verify it on a pool type it reads; concentrated pools (Orca Whirlpool, Raydium CLMM, Meteora DLMM and similar) are marked as not applicable, not as burned.
• Grades move. PVE went from F to D on the same day when its liquidity flag changed. BONK went from A / 90 to B / 82 later on 10 November 2026 when TOP1_LARGE appeared. Refresh before you rely on a screenshot.
• A high grade is not a buy rating. A low grade on a well-known mint is not, by itself, proof of fraud.

Warning signs, summarised

FAQ

What does a Solana token risk score mean?

On Rugsight it is an automated 0–100 technical score of public on-chain and market data: mint and freeze authority, Token-2022 controls, real-wallet holder concentration, liquidity and trading age. The letter grade is just that number in bands (A is 85 or above, B 70, C 50, D 30, F below 30). It is not financial advice and it is not a claim that the token is safe or is a proven rug.

What is the difference between grade A, B, C, D and F?

A (85–100) is "low risk on the checks we run". B (70–84) is moderate risk — read the flags. C (50–69) is elevated risk with several warning signs. D (30–49) is high risk with serious warning signs. F (0–29) is extreme risk with strong rug-pull indicators. The verdict text on each token page uses those phrases. The flags under the grade tell you which checks fired.

Is a grade A Solana token safe to buy?

No score we publish, including an A, means a token is safe to buy. Official TRUMP scored A / 100 on 10 November 2026 because no warning flags fired. BONK was A / 90 earlier that day and B / 82 by 23:18 America/Fortaleza once TOP1_LARGE appeared alongside TOP10_ELEVATED. That does not forecast price, rule out a later dump, or replace reading the mint, the pool and the flags. A new token can also reach 85 (an A) if only VERY_NEW applies. See the live BONK page for the current figure.

Why does a well-known token like JUP or RENDER get a C or D?

Because the grade only reflects the checks above, not how famous the ticker is. On 10 November 2026, Jupiter (JUP) scored C / 65 from TOP10_HIGH and TOP1_WHALE. RENDER scored D / 32 from active mint and freeze authorities plus holder flags. Those are technical findings. They are not a verdict that the project is a scam, and they are not hidden just because the name is familiar.

What does grade IC mean on a Rugsight score?

IC means issuer-controlled. The mint is on a short allowlist and its live authorities still match the issuer addresses we pinned (for example Circle on USDC, Coinbase on cbBTC). The letter A–F is replaced so a designed mint/freeze control is not shown as a rug pattern, but the technical score still deducts for those controls (USDC was IC / 50, cbBTC IC / 15 on this check). See what issuer-controlled means.

Why can the score change during the same day?

The inputs can change: liquidity can move between buckets, a holder flag can appear or drop, or a cache refresh can replace an older snapshot. PVE was F / 20 earlier on 10 November 2026 and D / 30 later the same day when the liquidity flag went from LIQUIDITY_VERY_LOW to LIQUIDITY_LOW. BONK was A / 90 earlier that day and B / 82 by 23:18 America/Fortaleza when the largest real wallet crossed 10% (TOP1_LARGE). Re-paste the mint rather than trusting an old screenshot.

Summary

Read the flags, then the number, then the letter. Rugsight's score is a 0–100 technical summary with fixed bands: A 85, B 70, C 50, D 30, F below that, and IC for allowlisted issuer-controlled mints. An A is not a buy signal; a C or D on a well-known token is not automatically a scam label; an IC is not an A. Paste the mint into Rugsight, compare related mints on a comparison page when you have two candidates, and use the full checklist for everything the score does not cover.

Rugsight is an automated technical tool. Nothing here is financial advice.