QPUStatus

IBM Marrakesh

Superconducting Transmon • 156 Qubits • Heavy-Hex Lattice • Heron r2 • TLS Mitigation • IBM Quantum System Two • Deployed Nov 2024 (QDC 2024)
IBM QUANTUM PLATFORM
LOAD
--
Pending Jobs
Awaiting telemetry...
RESERVED ACCESS
SOON
--
Not Yet Tracked
IBM offers dedicated reservation windows for enterprise clients via IBM Quantum Premium Plan. Reserved session telemetry is planned for future integration.
ON-PREMISES
N/A
--
Not Applicable
ibm_marrakesh is a cloud-access system hosted within an IBM Quantum Data Center. It is not available as a standalone on-premises deployment for third parties.

Live Status: Currently, the IBM Marrakesh is Online via IBM Quantum Platform (0 jobs). Updated real-time for IBM Quantum circuit monitoring.

QPUStatus is independent. Data from provider APIs may vary from internal states. Trademarks property of IBM Quantum. Not affiliated.

Live Network Load

IBM Quantum Platform

*Metric: Total number of jobs pending execution (Queue Depth) on ibm_marrakesh via IBM Quantum Platform.

Reserved Access

Reserved Session Telemetry Coming Soon

On-Premises

Not Applicable for ibm_marrakesh

System Availability Trends

IBM Quantum Platform Availability 100%
30 days ago Today

Detailed Connectivity (Last 7 Days)

Unofficial Telemetry Dashboard

This is an independent tracking project. QPUStatus is not affiliated with, endorsed by, or partnered with IBM or IBM Quantum. Our data is gathered automatically via public API routing endpoints and may not perfectly reflect internal hardware states. For authoritative information, visit the IBM Quantum Platform.

Hardware Deep Dive

IBM Marrakesh (ibm_marrakesh) is a Heron r2 processor system announced at the IBM Quantum Developer Conference in November 2024, where IBM's Robert Chow described it as one of the newest additions to the Heron fleet. Like ibm_fez, it runs the 156-qubit Heron r2 processor with tunable couplers and TLS mitigation. IBM names its QPUs after cities where the company has offices or operations around the world; ibm_marrakesh and ibm_fez are both named after Moroccan cities, though neither name is intended to indicate the physical data center location. As multiple Heron r2 systems exist within the IBM Quantum fleet, ibm_marrakesh provides additional capacity for users whose jobs would otherwise queue behind ibm_fez. IBM's QDC 2025 benchmarking report noted that across the Heron fleet, systems had reached 330,000 CLOPS through software improvements, compared to 200,000 at the end of 2024 and 150,000 at the QDC 2024 launch.

Technical Specifications

Architecture Fixed-Frequency Superconducting Transmon
Processor Family Heron r2 Source
Physical Qubits 156 Source
Topology Heavy-Hexagonal Lattice (max 3 neighbors per qubit)
Native Gates CZ, ID, RZ, SX, X Source
Tunable Couplers Yes — suppresses crosstalk to <0.1% level
TLS Mitigation Yes — controls the chip's TLS environment for improved coherence stability Source
Median 2Q Gate Error ~3×10−3 (best edge: ~1×10−3) Source
Max Two-Qubit Gates 5,000 (verified at 100-qubit, depth-100 circuit scale) Source
CLOPS 150,000–200,000 (at QDC 2024 deployment); 330,000 reported for the Heron fleet at QDC 2025 Source
Median T1 / T2 ~100–200 µs / ~50–150 µs Source
Operating Temperature ~15 millikelvin (dilution refrigerator)
System IBM Quantum System Two
Introduced November 2024 (IBM Quantum Developer Conference 2024) Source
Cloud Access IBM Quantum Platform (IBM Cloud) via Qiskit SDK

Common Provider Questions

ibm_marrakesh and ibm_fez use the same processor generation. How should I choose between them?

Both systems run the Heron r2 processor and share the same native gate set, topology, and general fidelity profile. In practice, the main reason to choose one over the other is queue length and current calibration state. IBM Quantum Platform shows live pending job counts and the most recent calibration data (T1, T2, per-edge 2Q error) for each system; submitting to whichever has a shorter queue or better current-day calibration numbers for your qubits of interest is the standard approach. Researchers targeting specific qubit subsets may also check the per-edge error map on the platform and select the system where that subgraph has the lowest reported errors on the day of their experiment, since calibration drifts over time on both systems independently.