IBM C1000-179 — Fundamentals of Quantum Computing Using Qiskit v2.X Developer Passing target: 24/34. All code outputs were verified with qiskit 2.5.1 / qiskit-ibm-runtime 0.48.0.

Answer Key

Q	Ans	Q	Ans	Q	Ans	Q	Ans
1	B	10	A	19	A	28	A
2	C	11	B	20	A	29	D
3	A	12	C	21	C	30	C
4	A	13	C	22	A	31	B
5	D	14	A	23	B	32	C
6	B	15	B	24	D	33	D
7	C	16	D	25	A, C	34	B
8	B	17	C	26	B
9	A	18	B	27	B

Explanations

Q1 — B (3). Layers: h(0) → cx(0,1) → the two measurements (they act on different qubit/clbit pairs, so they share one layer). Barriers are directives and do not add to depth(). Distractors: C (4) counts the barrier as a layer; D (5) additionally serializes the two measurements; A forgets the measurement layer entirely. Verified by execution.

Q2 — C (sdg). inverse() replaces each gate with its inverse in reverse order; S† is sdg. S is unitary, so it certainly has an inverse (D false); S·S = Z but S⁻¹ ≠ Z (B is the trap for people who remember S² = Z). Verified by execution.

Q3 — A. Classical feedforward in Qiskit 2.x uses the context-manager form with qc.if_test((cr, 1)):. Distractors: B (c_if) is the deprecated pre-2.x API, removed in Qiskit 2.0; C is not a real signature; D is ordinary Python evaluated at build time — cr has no runtime value there.

Q4 — A (1.0). List binding assigns values in the order of qc.parameters, which is sorted alphabetically by name — not gate-insertion order. So a → 1.0 and b → 2.0, and rx(a) becomes rx(1.0). Distractors: B (2.0) assumes insertion order (ry was added first); D is false — lists are accepted. Verified by execution: ry [2.0], rx [1.0].

Q5 — D ({'100': 0.5, '101': 0.5}). a.tensor(b) places b's qubits as the low qubits (0–1) and a's qubit on top (qubit 2). X sits on qubit 2 (leftmost bit = 1), H on qubit 0 (rightmost bit is 0/1 with prob ½ each). Distractor A puts the operands the other way around (b.tensor(a) behavior); B forgets the H superposition. Verified by execution.

Q6 — B. generate_preset_pass_manager(optimization_level=..., backend=...) then pm.run(qc) is the v2.x path to an ISA circuit. A: QuantumCircuit has no .transpile() method. C: qiskit.execute was removed in Qiskit 1.0. D: decompose() unrolls definitions but knows nothing about a backend's basis or coupling map.

Q7 — C ((-1+0j)). Little-endian twice over: in the label '01' the rightmost character is qubit 0 (so qubit 0 = |1⟩), and in the Pauli string 'IZ' the rightmost character acts on qubit 0. Z on |1⟩ gives eigenvalue −1. Distractor A (+1) is the big-endian reading (Z on qubit 1, which is |0⟩). Verified by execution.

Q8 — B (False True). Single X and Z anticommute. But XX vs ZZ: each position contributes a −1, and (−1)² = +1, so the two-qubit operators commute. Distractor A assumes anticommutation is inherited termwise; C forgets X and Z anticommute at all. Verified by execution.

Q9 — A. After H+CX the state is (|00⟩+|11⟩)/√2; x(0) flips qubit 0 (the rightmost bit) in both branches: |01⟩ and |10⟩, ½ each. B ignores the final X; C/D invent a collapse that never happened. Verified by execution.

Q10 — A (True). H·X·H = Z is a standard conjugation identity, and equiv compares up to global phase. Operator(Pauli('Z')) is perfectly legal (D false). Verified by execution.

Q11 — B (['IZI']). from_sparse_list([("Z", [1], 2.0)], num_qubits=3) puts Z on qubit 1. With little-endian printing, qubit 0 is rightmost, so qubit 1 is the middle character: IZI. Distractors A/C place it on qubit 2 or qubit 0 (C is the "index = position from left" misread). Verified by execution.

Q12 — C (0.5). Fidelity between pure states is |⟨+|0⟩|² = (1/√2)² = 0.5. Distractor B (0.707) is the overlap amplitude, forgetting to square. Verified by execution.

Q13 — C (5). FakeManilaV2 mimics the 5-qubit ibmq_manila device. Verified by execution.

Q14 — A. Open the session as a context manager and pass it to the primitive with mode=session. B passes the session as a PUB; C is V1 Sampler syntax; D nests objects incorrectly (SamplerV2 takes mode=, not backend=).

Q15 — B. Independent circuits with no feedback → Batch mode: jobs are scheduled together efficiently without reserving exclusive iterative access. Session (A) is for feedback loops; 200 separate jobs (C) re-queue every time.

Q16 — D (2048). Shots precedence in V2: per-PUB shots beat the run(..., shots=...) value, which beats options.default_shots. The PUB (qc, None, 2048) wins over shots=100. Distractors: A inverts the precedence; C — specifying both is legal, not an error. Verified by execution.

Q17 — C. V2 primitives on IBM hardware validate that circuits match the backend target (ISA). Raw h/cx circuits are rejected — the service does not auto-transpile (A is pre-primitives behavior).

Q18 — B ({'1': 100}). Sampler data is keyed per classical register name: res.data.a is the 1-bit register holding qubit 0's measurement, and x(0) makes it 1. Distractors: A forgets the X; C is the concatenated two-register string (that's what join_data() would give); D — .data.meas only exists when measure_all() created a register named meas. Verified by execution.

Q19 — A (8 1). get_bitstrings() returns one string per shot → 8 of them; num_bits is the width of the register → 1 (one qubit measured into the auto-created 1-bit meas register). B swaps the two numbers. Verified by execution.

Q20 — A. DD is configured on the options tree: options.dynamical_decoupling.enable = True plus sequence_type = "XY4". B (resilience_level) doesn't exist on Sampler; C uses the removed V1 set_options; D — run() takes PUBs, not option kwargs.

Q21 — C. An ideal Bell state yields only '00' and '11', each ≈ 50% of 1024 shots with statistical fluctuation. A has no randomness, B is the uniform 4-outcome distribution of h(0); h(1), D is the |Ψ+⟩ signature (that circuit would need an extra X).

Q22 — A (1.0). For the Bell state, the qubits are perfectly correlated in the Z basis, so ⟨ZZ⟩ = 1. B (0.0) is ⟨ZI⟩ or ⟨IZ⟩ — single-qubit Z on a Bell state is 0, the classic confusion. Verified by execution (0.9999999999999998).

Q23 — B ([ 1. -1.]). ry(0)|0⟩ = |0⟩ → ⟨Z⟩ = +1; ry(π)|0⟩ = |1⟩ → ⟨Z⟩ = −1, in the order the parameter sets were given. A reverses the order; C would need θ = π/2. Verified by execution.

Q24 — D. Two things must both happen: the circuit is the ISA circuit isa, and the observable is remapped with obs.apply_layout(isa.layout). Distractors: A is the top exam trap — transpilation may permute qubits, and nothing remaps the observable for you; B submits the untranspiled circuit; C is the V1 Estimator API.

Q25 — A and C. Dynamical decoupling and Pauli twirling act on the circuit before/during execution → suppression. ZNE (B) and PEC (D) are classical post-processing of results → mitigation.

Q26 — B. h(1) puts the H box on the lower wire (q1), and cx(1, 0) puts the control dot on q1 with the ⊕ target on q0. Distractor A is the mirror-image h(0); cx(0, 1); D reverses time order. Drawing verified by execution.

Q27 — B. Dominant 00/11 peaks with small 01/10 leakage is a Bell state plus readout error. A would give four roughly equal bars; C a single 11 bar; D would peak at 01/10 instead.

Q28 — A. Only qubit 0 gets an H, so it points along +X (|+⟩); qubit 1 stays |0⟩ = +Z. There is no entanglement, so no zero-length vectors (D describes a Bell pair). B is the endianness swap — remember the first subplot is qubit 0.

Q29 — D. plot_state_city is the 3D "cityscape" of Re(ρ) and Im(ρ). Qsphere shows amplitudes/phases on a sphere, paulivec shows Pauli-basis expectation bars, and bloch_multivector shows per-qubit arrows.

Q30 — C (0.6). ZZ eigenvalue is +1 for 00/11 and −1 for 01/10: (400 + 400 − 100 − 100)/1000 = 0.6. Distractor A (0.8) only sums the positive-parity counts; D flips the sign.

Q31 — B. V2 primitive execution is asynchronous: run() returns a job handle immediately, and job.result() blocks until completion. get_counts (C) is the old Result API; status() (D) is a passive query.

Q32 — C (3). A PrimitiveResult is iterable with one PubResult per PUB submitted — three circuits → len(result) is 3. Distractors B/D confuse the number of PUBs with shots. Verified by execution.

Q33 — D. OpenQASM 3 assigns measurements: c[0] = measure q[0];. Distractors: A is QASM 2's arrow syntax; B (creg) is a QASM 2 declaration (QASM 3 uses bit[2] c;); C has the type backwards — QASM 3 declares qubit[2] q;, and option C's qubit q[2] puts the size on the wrong side. Verified: the dumped program contains bit[2] c;, qubit[2] q;, c[0] = measure q[0];.

Q34 — B ({'10': 100}). x q[0] sets qubit 0 to 1, but the measurements are crossed: q0 → c[1] and q1 → c[0]. So c1 = 1, c0 = 0, and the printed string is c1c0 = '10'. Distractor A ('01') follows the qubit values instead of the classical-bit mapping; D forgets the X entirely. Verified by execution.