Entanglement enables coherence transfer and feedforward corrects the state in quantum teleportation on superconducting dynamic circuits

Abstract Quantum teleportation transfers an unknown qubit state from a sender to a receiver using a shared entangled pair and a classically conditioned correction without moving the physical carrier. On the dynamic-circuit hardware, the complete protocol (which includes mid-circuit measurement and conditional feedforward) now runs on a single circuit. We use this to benchmark three-qubit teleportation on a 156-qubit IBM Heron r2 processor (ibm_fez). A single-qubit RY(θ) input was teleported at six angles along a meridian of the Bloch sphere and reconstructed by state tomography alongside two ablation controls; one without entanglement and the other without feedforward. The full protocol returns a mean fidelity of about 0.945, nearly constant across the angles and above the 2/3 full-Bloch-sphere reference at all six sampled angles. Removing feedforward lowered the fidelity to about 0.51, close to a maximally mixed state. On the other hand, removing entanglement yields (1 + cos²θ)/2 with a mean near 0.75, transferring the population but losing the coherence. Together, these results establish that entanglement and feedforward are both necessary. The calibration-limited fidelity serves as a reference for this class of hardware. A separate run returned a Bell-state fidelity value of 0.957. A branch-resolved analysis confirmed that the four measurement branches are Pauli rotations of the target realigned by feedforward. The reported fidelity depends only on the rotation angle and the device noise.