QC Design’s Meridian Cuts Logical Error Rates by Double-Digits with Purpose-Built Architecture System

# What Meridian is and why it matters QC Design, a quantum software startup based in Ulm, released white paper results for Meridian, a platform that automates and optimizes architectures for fault-tolerant quantum computing (FTQC). Meridian is framed as a purpose-built system that searches across the many interacting design layers that determine logical error rates: quantum error-correction (QEC) code selection, syndrome-extraction circuit design, physical routing and layout, compilation choices, and low-level control sequences.

# How Meridian works Meridian combines specialized AI agents with Plaquette, QC Design's design-automation simulation environment described as an independent ''world model.'' Plaquette simulates hardware-level noise channels including dephasing, crosstalk, leakage, and measurement error across several hardware modalities: superconducting qubits, silicon spins, neutral atoms, trapped ions, and photonics. Candidate architectures proposed by Meridian are validated against these physical noise models before being accepted.

# Benchmark scope and results The evaluation suite covered more than 100 fault-tolerance design tasks spanning 10 QEC code families (including rotated planar and color codes) and varied connectivity classes (square and hexagonal grids, narrow ribbons, all-to-all couplings). Key headline results in the white paper include:

• Enabling deeper logical circuits by a median factor of 14.6× according to the benchmarking metrics.

# Concrete case study: silicon-spin hardware One highlighted example used a silicon-spin platform and a distance-5 color code logical memory experiment. Meridian autonomously repurposed unallocated lattice sites as helper qubits to repeatedly transfer data-qubit state information and reset leakage channels. That tactic produced a reported 29-fold reduction in logical error rates compared with the literature baseline algorithm for that task.

# Why Plaquette matters

# Implications for hardware developers QC Design frames Meridian as a tool to reduce physical-qubit overheads and shorten the path to fault-tolerant operation. By automating multi-layer architectural choices, Meridian aims to deliver tailored blueprints for different hardware modalities and topologies, which the company suggests could accelerate development timelines and lower capital costs.

# What the benchmarks do and don't say The published results are reported in a white paper and include direct comparisons to prior literature methods and to a general-purpose GPT-6 Astra agent on the same evaluation harness. The white paper provides task-level results, multiple hardware modalities, and at least one concrete hardware-case demonstration with an explicit mitigation strategy (helper-qubit leakage resets).