Logical error rate (Logical error)
How often an error-corrected (logical) qubit still fails, usually per round of error correction.
When results can be compared
Depends on the code, its size (distance), what counts as one cycle, and whether failed runs are discarded. Compare only identical set-ups.
Where it comes from
Standard measure in quantum error correction.
- Type
- Measurement, system level
- Unit
- error per cycle
- Better is
- Lower
- Applies to
- machines running error correction
Results
Set-up: concatenated iceberg d4 memory. Only these results are compared with each other.
| Machine |
Value | Set-up | Published | Measured by | Grade | Record |
| Quantinuum Helios |
4 × 10⁻⁵ |
concatenated iceberg d4 memory |
25 Feb 2026 |
Maker |
B |
r-000019 |
Set-up: surface code d7 memory. Only these results are compared with each other.
| Machine |
Value | Set-up | Published | Measured by | Grade | Record |
| Google Willow |
7.72 × 10⁻⁴ |
surface code d7 memory |
11 Nov 2025 |
Maker |
B |
r-000002 |
| Google Willow |
1.43 × 10⁻³ |
surface code d7 memory |
9 Dec 2024 |
Maker |
B |
r-000001 |
Set-up: color code d5 memory. Only these results are compared with each other.
| Machine |
Value | Set-up | Published | Measured by | Grade | Record |
| Google Willow |
0.011 |
color code d5 memory |
26 May 2025 |
Maker |
B |
r-000003 |
Set-up: carbon code bell postselected. Only these results are compared with each other.
| Machine |
Value | Set-up | Published | Measured by | Grade | Record |
| Quantinuum H2 |
1 × 10⁻⁵ |
carbon code bell postselected |
2 Apr 2024 |
Maker |
B |
r-000024 |
