IBM Quantum Campaign · Marrakesh · Boston · Kingston · May-August 2026

We do not ask you to take our conclusions on faith.

We publish a selection of original documents exported from the IBM Quantum Platform. Each evidence record preserves its Job ID, QPU, execution details, circuits, results, and the SHA-256 of the published PDF, allowing any third party to verify the exact file being reviewed.

Explore all 25 evidence records Download SHA256SUMS Manifest JSON
25selected IBM workloads
3physical QPUs in the campaign
155/156visible positions on Boston
576PUBs in one Boston workload
98.291%4026/4096 against the documented target
Editorial principle. This page is not a screenshot gallery or a marketing demo. The selection allows researchers, engineers, technical investors, and innovation leaders to reconstruct a complete experimental record: functional calibration, package selection, physical expansion, near-full-chip execution, phase and interaction gates, statistical validation, PUB scale, and operational repeatability.

Five starting points

If you open only five documents, start here. Each card links to the complete workload record and the SHA-256 of the published PDF.

EVIDENCE 09 · BOSTON

155/156: the strongest visual evidence of physical expansion

The visible physical footprint reaches 155 indices on a 156-qubit processor. Formally, 155 qubits define 2^155 ≈ 4.57×10^46 dimensions. The milestone is not to enumerate them, but to construct and execute a parameterized trajectory across virtually the entire available physical resource.
EVIDENCE 19 · KINGSTON

4026/4096 = 98.291%: the success anchor against the target

With the QPE target fixed before IBM returned the result, 4026 matches are not a peak selected after the fact; they validate a prior prediction. On that basis, the success rate against the target is 98.291015625%.
EVIDENCE 11 · BOSTON

Backend calibration + 1012/1024: NISQ context in the same record

This avoids treating the QPU as ideal. The reader can see the imperfect hardware state and, in the same time context, the resulting distribution.
EVIDENCE 12 · BOSTON

576 PUBs: orchestration at scale

Direct evidence of the ability to construct and deliver hundreds of quantum work units within a single workload.
EVIDENCE 21 · MARRAKESH

QPE-V88: 150/156 visible on Marrakesh

Near-full-chip selection is not exclusive to Boston. V88 reproduces a 150-visible-qubit footprint on a second QPU.

The 25 evidence records

Filter by QPU or evidence type. Open any record to see exactly what IBM prints, what it means, what to verify, and the precise boundary of the claim.

Functional calibration
01
Campaign launch: real, immediate, traceable physical executiond84ertmgbeec73ao1cg0 · 2026-05-17
Marrakesh
PUBs18
Shots512 default
StatusCompleted
Pending0 s
QPU usage5 s
Total8 s

What the PDF contains

The first page identifies qpe-test v8, the ibm_marrakesh backend, Sampler, and 18 PUBs. IBM marks the job Completed, with 0 s Pending, 5 s Usage, and 8 s total. The PDF also preserves measurement results and access to the circuit/Qiskit.

What it means

This is the operational baseline. Before hundreds of PUBs or near-full-chip runs, it shows that the pipeline was already sending real workloads to a physical QPU and receiving results within seconds.
What the reader should verify
Verify ibm_marrakesh, 18 PUBs, Completed, 0 s Pending, 5 s Usage, and default_shots=512.
Claim boundary: It does not independently demonstrate fidelity optimization or scaling; its value is as a baseline.
SHA-256 · cfd13e32d0464a7e010b551be7e98c4bf4645c901441d54b71f50e0ebf32e95a
Open IBM PDF View manifest
02
Early scaling: more PUBs and shots without losing immediate accessd84f03voha1c73bnnu80 · 2026-05-17
Marrakesh
PUBs22
Shots1024 default
StatusCompleted
Pending0 s
QPU usage8 s
Total14 s

What the PDF contains

IBM records 22 PUBs, 1024 default shots, Completed status, 0 s Pending, 8 s QPU Usage, and 14 s total.

What it means

The initial phase was not limited to microjobs. The system increases both work units and sampling while maintaining a clean operational route to the QPU.
What the reader should verify
Review # of PUBs, default_shots, Pending, Usage, and Completed.
Claim boundary: PUBs and qubits are different dimensions; 22 PUBs does not mean 22 qubits.
SHA-256 · eb617a70a1992dc7ad632dc7067b81fc0bfee407504266a332be1eca9e6d0df1
Open IBM PDF View manifest
03
Deliberate microprobe: measure before scalingd8526nftjchs73bqqdg0 · 2026-05-17
Marrakesh
PUBs1
Shots512
StatusCompleted
Pending0 s
QPU usage2 s
Twirlingmeasurement enabled

What the PDF contains

The workload drops to 1 PUB, retains 512 shots, and shows measurement twirling enabled. IBM records 0 s Pending and 2 s Usage.

What it means

In a functional calibration campaign, a microprobe can test a region, preparation, or configuration at low cost before reusing what was learned.
What the reader should verify
Verify 1 PUB, 512 shots, twirling.enable_measure=true, 0 s Pending, and 2 s Usage.
Claim boundary: The PDF proves configuration and execution; the detailed scientific intent belongs to the internal protocol.
SHA-256 · 3899b565eb398a4c37d6932252433a20a74357e821698a48c8d36238afa325ce
Open IBM PDF View manifest
04
14-qubit package: visible phase parameterizationd852p4o0bvlc73d54u70 · 2026-05-17
Marrakesh
PUBs1
Shots1024
StatusCompleted
Pending0 s
QPU usage2 s
Visible registerc14 / q[1]-q[14]

What the PDF contains

The circuit shows q[1] through q[14], register c14, numerous RZ rotations with different angles, and X layers.

What it means

This shows the physical language of QPE/ECHELEN: not an isolated RZ, but an ordered vector of distinct parameters on a specific physical package. Formally, 14 qubits define 2^14 possible dimensions; operationally, preparation and gates determine the executed trajectory.
What the reader should verify
Locate q[1]…q[14] and RZ values such as 0.30128, -π, 2.45223, and -0.22583, interleaved with X.
Claim boundary: The circuit does not enumerate every basis state in Hilbert space.
SHA-256 · 25bc39c447b0a8fdca64adbd1ff8794441f03350d2d3e1cb0795c0bc91b17156
Open IBM PDF View manifest
05
Distributed physical selection: a package need not be contiguousd8526u7tjchs73bqqdn0 · 2026-05-17
Marrakesh
PUBs1
Shots512
StatusCompleted
Pending0 s
QPU usage2 s
Visible qubits91, 92, 98, 109, 110, 111, 118, 129

What the PDF contains

The circuit uses a distributed set of physical qubits and applies RZ/X sequences with individual parameters.

What it means

This is intuitive evidence of physical navigation: selection can move across chip regions instead of expanding mechanically as q[0]…q[n].
What the reader should verify
Observe the widely separated physical indices and the distinct RZ parameterization on each line.
Claim boundary: The PDF shows the selection; the index-selection algorithm is Human Blooming IP.
SHA-256 · af33321a3e28d48b733210af6f3ba3151bac11e0f3f959368b0fd85202af4fee
Open IBM PDF View manifest
Boston: expansion and scale
06
Changing QPUs: the pipeline moves from Marrakesh to Bostond8bek5ce6kns73aav4cg · 2026-05-27
Boston
PUBs12
StatusCompleted
Pending0 s
QPU usage4 s
Total6 s
ProgramSampler

What the PDF contains

IBM identifies ibm_boston, 12 PUBs, Sampler, Completed, 0 s Pending, 4 s Usage, and 6 s total.

What it means

The value is architectural: the layer is no longer tied to one QPU and retains its execution model when the backend changes.
What the reader should verify
Verify “Quantum computer: ibm_boston,” 12 PUBs, Completed, 0 s Pending, and 4 s Usage.
Claim boundary: This does not prove complete provider independence; it does demonstrate a multi-QPU campaign.
SHA-256 · dd8fba678434116f9cec395a614ccd7d08578cd9bffb09d2cdc513956f1329e9
Open IBM PDF View manifest
07
Boston reaches q[149]: physical expansion before 155/156d8gckik2upec739jiopg · 2026-06-04
Boston
PUBs4
StatusCompleted
Pending0 s
QPU usage2 s
Total7 s
Visible footprintthrough q[149]

What the PDF contains

On the final pages, the circuit reaches q[149] and shows long RZ(π/2) and X layers.

What it means

This is the transition toward near-full-chip. A width of 150 qubits formally corresponds to a Hilbert-space dimension of 2^150 ≈ 1.43×10^45.
What the reader should verify
Find q[131]…q[149] and the repeated RZ/X sequence; verify 4 PUBs and 2 s Usage on the cover.
Claim boundary: The presence of 150 indices does not by itself demonstrate maximal entanglement across all of them.
SHA-256 · 898014eb20d0a60bf8f44ccc931b56ba1e5c247dbb5bba2c5e0a2a0d46cc4454
Open IBM PDF View manifest
08
Highly concentrated result with a non-trivial circuitd8jh82u6983c73dtcg30 · 2026-06-08
Boston
PUBs3
StatusCompleted
Pending0 s
QPU usage4 s
Visible result388/400
Another PUB2048 shots

What the PDF contains

The main view shows 388/400. The PDF also contains a 2048-shot PUB and a wide circuit with numerous RZ values near -π and X layers.

What it means

It combines a concentrated distribution with a complex circuit structure; this is not an empty job designed merely to produce a low runtime.
What the reader should verify
Verify 388/400, then review the 2048-shot PUB and its circuit.
Claim boundary: Fidelity against a target should be claimed only when the target was fixed in advance.
SHA-256 · 916ca4c603663245cad3d320a81c46a04ccda3fba4c9d137034be154689ba21e
Open IBM PDF View manifest
09
155/156: the strongest visual evidence of physical expansiond8jfoj42upec739ng600 · 2026-06-08
Boston
PUBs32
Shots512 default
StatusCompleted
Pending0 s
QPU usage8 s
Footprintq[0]-q[154]

What the PDF contains

The circuit spans eight pages, tracing q[0]-q[25], q[26]-q[51], intermediate regions, and finally q[131]-q[154]. Each block contains distinct RZ values and X layers.

What it means

The visible physical footprint reaches 155 indices on a 156-qubit processor. Formally, 155 qubits define 2^155 ≈ 4.57×10^46 dimensions. The milestone is constructing and executing a parameterized trajectory over virtually the entire physical resource.
What the reader should verify
Review the complete PDF and follow indices q[0] through q[154]. Observe variation in the RZ parameters.
Claim boundary: 155 visible indices do not by themselves equal 155 maximally entangled qubits.
SHA-256 · 7b625c698785425f8dfefb5337967d4051517ae408b0598069ff0205016eed49
Open IBM PDF View manifest
10
312 PUBs: workload volume is not circuit widthd8h2o1pe8nrc73bgmg60 · 2026-06-05
Boston
PUBs312
StatusCompleted
Pending241m 18s
QPU usage1m 25s
Total246m 28s
PUB shown1024 shots / q[43]-q[44]

What the PDF contains

IBM records 312 PUBs and 1m25s of QPU Usage. One visible PUB shows 1024 shots and a small two-qubit circuit.

What it means

This demonstrates a second scaling dimension: batching. A workload can contain hundreds of work units even when each individual PUB is small.
What the reader should verify
Compare the 312-PUB record with the displayed q[43]-q[44] PUB.
Claim boundary: 312 PUBs does not mean 312 qubits.
SHA-256 · 707130dfb5e4fc2ad01771751a017aa1375e618672bf77d3cf5d6d5c174e5962
Open IBM PDF View manifest
Boston: calibration and gates
11
Backend calibration + 1012/1024: NISQ context in the same recordd9sc8uvpemts73ctic0g · 2026-08-09
Boston
PUBs56
StatusCompleted
Pending1 s
QPU usage17 s
Result1012/1024 = 98.828 %
Calibrationreadout 4.15E-3 / CZ 1.624E-3

What the PDF contains

The workload records 56 PUBs and 17 s. The supplementary calibration PDF fixes the backend state at job start: median readout assignment error 4.15×10^-3 and median CZ error 1.624×10^-3.

What it means

This avoids treating the QPU as ideal. The reader can see the imperfect hardware state and, within the same time context, the resulting distribution.
What the reader should verify
Open both the workload and calibration capture; verify 1012/1024 and the median readout/CZ values.
Claim boundary: The physical calibration shown is IBM data; Human Blooming performs functional calibration and selection against that reality.
SHA-256 · 03c3bd682fba7fdd74d0cef20d46c7d1c38658bec103cbb0128a9dce47d8f92e
Open IBM PDF Open supplementary calibration View manifest
12
576 PUBs: orchestration at scaled9u3nhj43mgs73esf21g · 2026-08-12
Boston
PUBs576
StatusCompleted
Pending1 s
QPU usage18 s
Total11m 56s
OptionsDD off / gate & measurement twirling off

What the PDF contains

IBM prints 576 PUBs and 18 s QPU Usage. Visible options show dynamical decoupling disabled and both gate and measurement twirling disabled.

What it means

Direct evidence of the capacity to construct and deliver hundreds of quantum work units in one workload.
What the reader should verify
Verify 576 PUBs, 18 s Usage, 1 s Pending, and the Input options flags.
Claim boundary: The 11m56s total is not effective QPU time; IBM records 18 s Usage.
SHA-256 · 20782ced37a8fcca2d258b1039347e3737e566db36c05a2878b9a3a03922ddda
Open IBM PDF View manifest
13
Parameterized RZZ: explicit two-qubit interactiond9uakfou5hac73agvsm0 · 2026-08-12
Boston
PUBs12
StatusCompleted
Pending149m 21s
QPU usage15 s
Visible gatesRZ, X, RZZ
Visible RZZ labels34 labels

What the PDF contains

In addition to long RZ and X sequences, the export contains 34 visible RZZ labels, including parameters 0.1092646313, 0.0847138806, 0.1105729848, 0.1055566374, and 0.1176448989.

What it means

RZ changes phase; RZZ introduces a parameterized interaction between pairs. The circuit moves beyond independent rotations and adds an explicit interaction layer.
What the reader should verify
Find “RZZ” on pages 3-5 and review the consecutive parameter series.
Claim boundary: 34 RZZ labels is a documentary count of visible labels, not an IBM-certified ISA gate count.
SHA-256 · 01d23888d0ba5ac430b13604a0da1c7a271f33472077cad7d408b580e11ec994
Open IBM PDF View manifest
Kingston: sweep and validation
14
Kingston: 396/400 during the short-sweep phased90dsc0pknjs73a28pf0 · 2026-06-28
Kingston
PUBs5
StatusCompleted
QPU usage8 s
Result396/400 = 99.000 %
Circuit5 qubits RZ/X
Shots in circuit shown4096

What the PDF contains

The results view shows 396 of 400 in the dominant outcome. The circuit uses five qubits with RZ/X layers.

What it means

This is a calibration/selection point in the Kingston family, preceding the longer 4096-shot validations.
What the reader should verify
Verify 396/400 and review q[0]-q[4] and the RZ parameters.
Claim boundary: It should be expressed as success against the outcome/target defined for that PUB.
SHA-256 · b968f89c72da63f033ffa1abeca672ebcd63c32c3d9a4e8cdccdaa9d506fee0a
Open IBM PDF View manifest
15
379/400: publishing the variation tood90eifopknjs73a29grg · 2026-06-28
Kingston
PUBs7
StatusCompleted
QPU usage10 s
Result379/400 = 94.750 %
Circuit5-qubit RZ/X family

What the PDF contains

IBM records 7 PUBs and 10 s Usage. The main view shows 379 of 400.

What it means

Publishing it alongside 396/400 shows the pack does not select only the best result; experimental variability is part of the evidence.
What the reader should verify
Compare structure and parameters with d90dsc0pknjs73a28pf0.
Claim boundary: Differences between jobs cannot be assigned to one cause without reconstructing all conditions.
SHA-256 · 00de80a66d32256a6bc9dcfd16f5b85b1b917ecf76f426701bcc80d8e8e1c21e
Open IBM PDF View manifest
16
11 PUBs, 378/400: the intermediate stepd90h7nhropqc738doal0 · 2026-06-28
Kingston
PUBs11
StatusCompleted
QPU usage14 s
Result378/400 = 94.500 %
Circuit5 qubits RZ/X
Visible shots4096

What the PDF contains

The workload rises to 11 PUBs and 14 s. Page two shows five qubits and specific RZ parameters.

What it means

It connects the 5-7 PUB jobs with the 23-PUB / 4096-shot validations.
What the reader should verify
Verify 11 PUBs, 14 s, and 378/400; review the five-qubit circuit.
Claim boundary: The visible 4096 shots correspond to the PUB shown on the circuit page.
SHA-256 · 85fd99e94e17777f738849aafdb601509bcd63a4853c08002579ebda6b6ac6df
Open IBM PDF View manifest
17
Controlled parameter perturbation: same form, new anglesd90hci06c68s73ai370g · 2026-06-28
Kingston
PUBs11
StatusCompleted
Pending0 s
QPU usage14 s
Result376/400
Parametersmodified RZ angles

What the PDF contains

The five-qubit RZ/X structure is retained while specific angles change, for example 1.795607→1.818115, 1.829562→1.811311, and 1.833008→1.827725.

What it means

This is direct evidence of a sweep through parameter space, not simple repetition of a fixed template.
What the reader should verify
Place this PDF beside d90h7nhropqc738doal0 and compare the initial RZ values.
Claim boundary: The mathematical rationale for each perturbation belongs to the proprietary pipeline and documentation.
SHA-256 · 44013b70d2f35927c5a49ac1dc1cba4da8aa65fd904ffcc682718dea2d1792c6
Open IBM PDF View manifest
18
4007/4096: Kingston’s first long validationd90dq1opknjs73a28mvg · 2026-06-28
Kingston
PUBs23
StatusCompleted
QPU usage27 s
Result4007/4096 = 97.827 %
Pending1m 29s

What the PDF contains

The main distribution shows 4007 of 4096 versus 89 for the other outcome.

What it means

4096 shots greatly reduce the risk of mistaking a small-sample fluctuation for a robust result. This is the first of three statistical anchors.
What the reader should verify
Verify 4007+89=4096 and 23 PUBs / 27 s Usage.
Claim boundary: Interpreting this as success against a target requires the associated precommit chain; the campaign fixed the target beforehand.
SHA-256 · 37658e0124cd79d48eddd5bae6f173d4053e1f1fa16166ef9b5deee82712d00e
Open IBM PDF View manifest
19
4026/4096 = 98.291%: the success anchor against the targetd90hc7emvj5c73ej70j0 · 2026-06-28
Kingston
PUBs23
StatusCompleted
Pending0 s
QPU usage27 s
Result4026/4096 = 98.291 %
Shots4096

What the PDF contains

IBM literally shows counts of 70 and 4026, totaling 4096.

What it means

Because the QPE target was fixed before IBM returned the result, 4026 matches validate a prior prediction rather than an after-the-fact selection.
What the reader should verify
Verify 4026/4096 in IBM and compare the precommit timestamp with the execution timestamp.
Claim boundary: The IBM PDF proves the distribution; Human Blooming retains the manifest/precommit that proves the target predates it.
SHA-256 · 37985019fbb367b326eb18e85bb620cd7b1973aa5be5416156b48404cd3a1677
Open IBM PDF View manifest
20
3972/4096: repetition, not a single captured90h7786c68s73ai319g · 2026-06-28
Kingston
PUBs23
StatusCompleted
QPU usage27 s
Result3972/4096 = 96.973 %
Pending5 s

What the PDF contains

IBM records 3972 and 124, totaling 4096, at the same scale of 23 PUBs and 27 s.

What it means

Its strength comes from publication alongside 4007/4096 and 4026/4096: three independent runs, the same scale, and strongly concentrated results.
What the reader should verify
Compare the three Job IDs, their dates, and Usage.
Claim boundary: Rates should only be aggregated when they represent exactly the same target and criterion; the campaign retains that documentation.
SHA-256 · 2cd3887b1e16f762935826325ee87d8a1069d5183a368eafac344437555de4c3
Open IBM PDF View manifest
Marrakesh V88
21
QPE-V88: 150/156 visible on Marrakeshd9uuhot0vrcc73boij80 · 2026-08-13
Marrakesh
PUBs22
StatusCompleted
QPU usage15 s
Visible shots2048
Footprintq[0]-q[149]
GenerationQPE-V88

What the PDF contains

The circuit spans eight pages, beginning at $[0] and ending at $[149], with large RZ(π/2), X, and parameterized RZ families in between.

What it means

Near-full-chip selection is not exclusive to Boston. V88 reproduces a 150-visible-qubit footprint on a second QPU.
What the reader should verify
Find $[0] on page 3 and $[149] on page 8; verify 22 PUBs and 15 s Usage.
Claim boundary: 150/156 describes visible physical indices; it is not an official IBM “full-chip” benchmark.
SHA-256 · 4ffc14c3e04210babeb2b81ca104989bdef51e1547646ce7f6c3fa12f51b9bec
Open IBM PDF View manifest
22
48 PUBs / 17 s: V88’s repeated blockd9v16qvo3ppc73ajmqqg · 2026-08-13
Marrakesh
PUBs48
StatusCompleted
QPU usage17 s
Pending70m 59s
Resultsvisible hexadecimal outcomes

What the PDF contains

IBM records 48 PUBs, 17 s Usage, and 70m59s Pending.

What it means

This belongs to a family of seven 48-PUB jobs with exactly 17 s QPU time, allowing quantum cost to be separated from queue behavior.
What the reader should verify
Compare # of PUBs=48 and Usage=17s with the other V88 members.
Claim boundary: Long Pending time is external scheduling, not quantum-computation cost.
SHA-256 · 217e347db97e7c27d50f50320b0f96f6fbd1a906a1c840fa03ee55b7e40362d9
Open IBM PDF View manifest
23
Queue control: same load and Usage, far longer waitd9v16sgb1g9c73a81umg · 2026-08-13
Marrakesh
PUBs48
StatusCompleted
QPU usage17 s
Pending92m 30s
Total93m 25s

What the PDF contains

The record has the same 48 PUBs and 17 s Usage as other block members, but 92m30s Pending.

What it means

This is methodological evidence: wall-clock time cannot directly measure algorithm cost when the scheduler introduces very different waits.
What the reader should verify
Compare it with d9v16qvo3ppc73ajmqqg: same PUBs and Usage, different Pending.
Claim boundary: This does not prove that QPE controls or eliminates IBM’s queue.
SHA-256 · f9e49f4611bb2968b74c21ed0f44e2ebfaad6784c4af028e0cdfbb3085fb4d1b
Open IBM PDF View manifest
24
168 PUBs: a volume jump within QPE-V88d9vgqqob1g9c73a8julg · 2026-08-14
Marrakesh
PUBs168
StatusCompleted
QPU usage34 s
Pending27m 25s
Resultswide hexadecimal outcomes

What the PDF contains

IBM records 168 PUBs, 34 s Usage, and Completed status.

What it means

V88 is not limited to repeating 22/40/48-PUB blocks; it markedly increases the volume of work units.
What the reader should verify
Verify 168 PUBs and 34 s Usage; observe the length of the outcomes.
Claim boundary: 168 PUBs are workload units, not 168 qubits.
SHA-256 · faf545fb9a967e7ee6961b106ab066da4d4fca9aa311db99a4222093193647b0
Open IBM PDF View manifest
25
V88 returns to a short execution: dynamic ranged9v623n2sl0c73blfmlg · 2026-08-14
Marrakesh
PUBs12
StatusCompleted
Pending1 s
QPU usage6 s
Total13 s
Resultswide hexadecimal outcomes

What the PDF contains

IBM records 12 PUBs, 1 s Pending, 6 s Usage, and 13 s total; the PDF preserves results and the circuit.

What it means

After 168 PUBs, the architecture can also return to small blocks. Maturity is not only scaling upward, but adapting the load.
What the reader should verify
Verify 12 PUBs, 6 s Usage, 1 s Pending, and 13 s total.
Claim boundary: This is one favorable-queue case; it does not imply all QPE jobs last 13 seconds.
SHA-256 · 67426cc2e9f38d470ce5b7f88dfdfbec366238ff0f75c6b05efeb69b73aee6c6
Open IBM PDF View manifest

Public integrity: what SHA-256 certifies

The Job ID identifies the execution on IBM. SHA-256 identifies the exact PDF published by Human Blooming. These are distinct layers. If even one byte changes, the SHA-256 will no longer match.

How to verify: download the PDF, calculate its SHA-256 locally, and compare it with the value shown in the record and manifest. All 25 hashes were calculated byte by byte and cross-checked using two independent implementations before publication.

SHA-256 Manifest

25 primary IBM Quantum PDFs plus 1 supplementary calibration PDF. CSV, JSON, and SHA256SUMS are also available for automated integration.

CSV JSON SHA256SUMS
# IBM Job ID QPU Date SHA-256 Source
01 d84ertmgbeec73ao1cg0 Marrakesh 2026-05-17 cfd13e32d0464a7e010b551be7e98c4bf4645c901441d54b71f50e0ebf32e95a PDF
02 d84f03voha1c73bnnu80 Marrakesh 2026-05-17 eb617a70a1992dc7ad632dc7067b81fc0bfee407504266a332be1eca9e6d0df1 PDF
03 d8526nftjchs73bqqdg0 Marrakesh 2026-05-17 3899b565eb398a4c37d6932252433a20a74357e821698a48c8d36238afa325ce PDF
04 d852p4o0bvlc73d54u70 Marrakesh 2026-05-17 25bc39c447b0a8fdca64adbd1ff8794441f03350d2d3e1cb0795c0bc91b17156 PDF
05 d8526u7tjchs73bqqdn0 Marrakesh 2026-05-17 af33321a3e28d48b733210af6f3ba3151bac11e0f3f959368b0fd85202af4fee PDF
06 d8bek5ce6kns73aav4cg Boston 2026-05-27 dd8fba678434116f9cec395a614ccd7d08578cd9bffb09d2cdc513956f1329e9 PDF
07 d8gckik2upec739jiopg Boston 2026-06-04 898014eb20d0a60bf8f44ccc931b56ba1e5c247dbb5bba2c5e0a2a0d46cc4454 PDF
08 d8jh82u6983c73dtcg30 Boston 2026-06-08 916ca4c603663245cad3d320a81c46a04ccda3fba4c9d137034be154689ba21e PDF
09 d8jfoj42upec739ng600 Boston 2026-06-08 7b625c698785425f8dfefb5337967d4051517ae408b0598069ff0205016eed49 PDF
10 d8h2o1pe8nrc73bgmg60 Boston 2026-06-05 707130dfb5e4fc2ad01771751a017aa1375e618672bf77d3cf5d6d5c174e5962 PDF
11 d9sc8uvpemts73ctic0g Boston 2026-08-09 03c3bd682fba7fdd74d0cef20d46c7d1c38658bec103cbb0128a9dce47d8f92e PDF
12 d9u3nhj43mgs73esf21g Boston 2026-08-12 20782ced37a8fcca2d258b1039347e3737e566db36c05a2878b9a3a03922ddda PDF
13 d9uakfou5hac73agvsm0 Boston 2026-08-12 01d23888d0ba5ac430b13604a0da1c7a271f33472077cad7d408b580e11ec994 PDF
14 d90dsc0pknjs73a28pf0 Kingston 2026-06-28 b968f89c72da63f033ffa1abeca672ebcd63c32c3d9a4e8cdccdaa9d506fee0a PDF
15 d90eifopknjs73a29grg Kingston 2026-06-28 00de80a66d32256a6bc9dcfd16f5b85b1b917ecf76f426701bcc80d8e8e1c21e PDF
16 d90h7nhropqc738doal0 Kingston 2026-06-28 85fd99e94e17777f738849aafdb601509bcd63a4853c08002579ebda6b6ac6df PDF
17 d90hci06c68s73ai370g Kingston 2026-06-28 44013b70d2f35927c5a49ac1dc1cba4da8aa65fd904ffcc682718dea2d1792c6 PDF
18 d90dq1opknjs73a28mvg Kingston 2026-06-28 37658e0124cd79d48eddd5bae6f173d4053e1f1fa16166ef9b5deee82712d00e PDF
19 d90hc7emvj5c73ej70j0 Kingston 2026-06-28 37985019fbb367b326eb18e85bb620cd7b1973aa5be5416156b48404cd3a1677 PDF
20 d90h7786c68s73ai319g Kingston 2026-06-28 2cd3887b1e16f762935826325ee87d8a1069d5183a368eafac344437555de4c3 PDF
21 d9uuhot0vrcc73boij80 Marrakesh 2026-08-13 4ffc14c3e04210babeb2b81ca104989bdef51e1547646ce7f6c3fa12f51b9bec PDF
22 d9v16qvo3ppc73ajmqqg Marrakesh 2026-08-13 217e347db97e7c27d50f50320b0f96f6fbd1a906a1c840fa03ee55b7e40362d9 PDF
23 d9v16sgb1g9c73a81umg Marrakesh 2026-08-13 f9e49f4611bb2968b74c21ed0f44e2ebfaad6784c4af028e0cdfbb3085fb4d1b PDF
24 d9vgqqob1g9c73a8julg Marrakesh 2026-08-14 faf545fb9a967e7ee6961b106ab066da4d4fca9aa311db99a4222093193647b0 PDF
25 d9v623n2sl0c73blfmlg Marrakesh 2026-08-14 67426cc2e9f38d470ce5b7f88dfdfbec366238ff0f75c6b05efeb69b73aee6c6 PDF
SUP d9sc8uvpemts73ctic0g Boston 2026-08-09 dbade2ac148da1975d908a1c04153c1909ed9c742626666f46cdb957c6e82a40 Open supplementary calibration

Evidence that completes the chain

Some important conclusions begin before IBM or end after IBM. They should therefore be audited together with their provenance artifacts, rather than forced into a single platform PDF.

Target / precommit

Manifest demonstrating that the Kingston target was fixed before the QPU result returned.

13/13 record

Thirteen commitments, timestamps, targets, results, and validation for every cycle.

C3

Manifest and aggregator for 99.26% / 1,181 errors, with the exact related Job IDs.

Blind transfer

QPU→GPU record, precommit, and the corresponding verifier result.

RTX 5090 + QPE

Independent third-party audit evidence covering the classical layer before the bridge.

Third-party audit reference hash:
fa86e04f7830545eb5795c3998aec2b29b3167163d7e6fa1761d6bc86e8aaf3e
This SHA-256 belongs to independently audited QPE evidence; it is not an IBM Job ID or a hexadecimal outcome.