MetaMask: software wallets rating breakdown
The most-used wallet in crypto, and the clearest illustration of what "free" actually costs. Credit where it is owed: its transaction simulation and phishing alerts are genuinely good, its audits are real and public, and it is non-custodial. It is not a bad wallet. It is an expensive one, in ways that are not on the price tag.
The most-used wallet in crypto, and the clearest illustration of what "free" actually costs. Credit where it is owed: its transaction simulation and phishing alerts are genuinely good, its audits are real and public, and it is non-custodial. It is not a bad wallet. It is an expensive one, in ways that are not on the price tag.
How the score is built
Each criterion is scored 0–10 and weighted. The median column is the middle score across every entry in this ranking, so a row reads as a position rather than a number.
| Criterion | What we measured | Weight | Score | Category median | Rank | Weighted gap |
|---|---|---|---|---|---|---|
| Security model and open code | — | 30% | 4 | 8 | 13 of 17 | -1.20 |
| Privacy | Infura — owned by Consensys, which owns MetaMask default RPC · sourced · 2026-07-14 · source | 20% | 2 | 5 | 16 of 17 | -0.60 |
| Transaction transparency | proprietary Consensys source-available — NOT open source since 2020 licence · sourced · 2026-07-14 · source | 15% | 8 | 5 | 3 of 17 | +0.45 |
| Swap cost | — | 15% | 3 | 5 | 11 of 17 | -0.30 |
| Recovery and portability | — | 10% | 9 | 9 | 1 of 17 | 0.00 |
| Networks | — | 10% | 9 | 7 | 1 of 17 | +0.20 |
Measured 14 July 2026 · weights and method · decided by security model and open code, worth -1.20 points against the median
Security model and open code: 4/10
Non-custodial, MPC or custodial — and where the key actually sits on your device. Then the licence, as it stands today rather than as folklore remembers it: several well-known wallets have quietly moved to source-available terms that are not open source. Audits count only when the firm is named and the report is public.
Scored 4 of 10 against a category median of 8, which places it 13th of 17 among software wallets on this criterion. At a 30% weight that is 1.20 points below the median contribution of the weighted total. The best score in the category is 10, the worst 2.
Privacy: 2/10
The most underrated question in this category: which RPC endpoint does the wallet talk to out of the box, and who owns it? If the default RPC is operated by the wallet’s own vendor, that vendor sees the IP address behind every balance check and every transaction, and can link it to your addresses. We read the default configuration and the privacy policy, and we quote the clause. Can you point it at your own node, and how hard is it?
Scored 2 of 10 against a category median of 5, which places it 16th of 17 among software wallets on this criterion. At a 20% weight that is 0.60 points below the median contribution of the weighted total. The best score in the category is 10, the worst 2.
Transaction transparency: 8/10
Does the wallet simulate the transaction and show you the actual outcome before you sign it, and does it warn you about known drainer contracts? The same question as clear signing on a hardware device, asked of software.
Scored 8 of 10 against a category median of 5, which places it 3rd of 17 among software wallets on this criterion. At a 15% weight that is 0.45 points above the median contribution of the weighted total. The best score in the category is 10, the worst 3.
Swap cost: 3/10
The built-in swap is how most free wallets are paid. We record the published service fee — and when a wallet does not publish one at all while still taking a cut of the spread, that silence is the finding and it is scored as one.
Scored 3 of 10 against a category median of 5, which places it 11th of 17 among software wallets on this criterion. At a 15% weight that is 0.30 points below the median contribution of the weighted total. The best score in the category is 10, the worst 1.
Recovery and portability: 9/10
Standard BIP39, and whether the seed restores in another wallet. A wallet you cannot leave is a custodian with extra steps.
Scored 9 of 10 against a category median of 9, which places it 1st of 17 among software wallets on this criterion. At a 10% weight that is exactly level with the median of the weighted total. The best score in the category is 9, the worst 3.
Networks: 9/10
Which chains genuinely work, from the wallet’s own documentation.
Scored 9 of 10 against a category median of 7, which places it 1st of 17 among software wallets on this criterion. At a 10% weight that is 0.20 points above the median contribution of the weighted total. The best score in the category is 9, the worst 2.
Other measurements
Its nearest neighbours in this ranking
| # | Entry | Security model and open code | How it differs |
|---|---|---|---|
| 10 | Rainbow | 8 | Ahead by 4 on security model and open code. |
| 11 | Keplr | 5 | Ahead by 1 on security model and open code. |
| 13 | Trust Wallet | 4 | Level on security model and open code; the gap is elsewhere. |
| 14 | Phantom | 2 | Behind by 2 on security model and open code. |
Questions about this score
How much does MetaMask charge for swaps?
+
0.875% of the swap amount, on top of the underlying exchange's fee and gas. Rabby charges 0.25% for the same function and Zerion 0.67%. On a $5,000 swap the difference between MetaMask and Rabby is about $31, and MetaMask's fee is disclosed in a support article rather than beside the quote.
Is MetaMask open source?
+
Not in the way most users assume. Its licence is more restrictive than the GPL and MIT licences used by Rabby, Rainbow and Frame, which limits forking and commercial reuse — the very scrutiny that makes 'open source' meaningful as a security claim.
Should I change MetaMask's default RPC?
+
It improves your privacy and it costs you features. The default is Infura, owned by MetaMask's parent, so that company sees the IP address behind your address queries. Pointing at your own node or another provider removes that — and disables Smart Transactions and MEV protection, which is a real trade rather than a free win.