What is QSB, and how does it work?
Quantum-Safe Bitcoin (QSB) is a way to spend bitcoin that a quantum computer cannot forge, using only the rules Bitcoin has today. StarkWare researcher Avihu Levy published the design on 9 April 2026, building on Robin Linus's Binohash.
- Lock the coins in a QSB output. It adds a second lock beside the usual signature, built on hash functions instead of elliptic curves.
- Search before spending. The spender runs a large GPU search offchain to solve a hash puzzle that binds the spend to one exact transaction.
- Send it straight to a miner. The transaction is larger than normal relay rules allow, so it goes to a miner directly. The first was mined on 26 Aug 2026.
QSB does not make Bitcoin as a whole quantum-safe or rescue coins whose public keys are already exposed, and each spend still needs about $59 of GPU compute and a direct path to a miner. StarkWare sees a soft fork as the better long-term answer.
Cost over time
GPU compute to prepare one transaction
Modelled estimates in USD
- 26 Aug 2026First QSB transaction$320
- 16 Sep 2026Challenge launched$320
- 23 Sep 202662 promoted improvements reported$67
- 27 Oct 2026Dashboard check · challenge closed$59
View the cost ledger and sources
| Date | GPU cost (USD) | Change vs first | Record | Source |
|---|---|---|---|---|
| 27 Oct 2026 | $59 | -81.5% | Dashboard check · challenge closed | Source ↗ |
| 23 Sep 2026 | $67 | -79.1% | 62 promoted improvements reported | Source ↗ |
| 16 Sep 2026 | $320 | 0.0% | Challenge launched | Source ↗ |
| 26 Aug 2026 | $320 | 0.0% | First QSB transaction | Source ↗ |
Which coins QSB protects
Protection depends on when the key is exposed
QSB vs a soft fork
| QSB | Soft fork · BIP-360 | |
|---|---|---|
| Consensus change | None | Required |
| Available today | Yes, since the first mainnet transaction on 26 Aug 2026 | No. BIP-360 is a proposal and has not been activated |
| Who is protected | Coins moved into QSB before their key is exposed | Holders who move to P2MR outputs: protection against long-exposure attacks. Short-exposure protection needs additional post-quantum signatures. |
| Cost to use | About $59 modelled GPU compute per transaction, plus fees | Fees depend on the spending script and transaction size. |
| How it is sent | Directly to a miner; nonstandard for public relay | Would need compatible wallets and nodes after activation |
| StarkWare’s view | An option that works today | The better long-term answer |
BIP-360 targets long-exposure attacks; it does not itself add post-quantum signatures. Read the proposal ↗
Credits
Avihu Levy published the QSB design on 9 April 2026. StarkWare, Yukon Research and Eigen Labs organized the optimization challenge.