Inspect the counts, then the method. These panels visualize recorded harness verdicts from immutable public receipts. They do not rank models or establish a general law of model behavior.
11 receipt files checked. Raw bytes, embedded digests, probe identities, and recorded totals are checked offline. This is not a signature verification, independent reproduction, or production approval.
Conditions differ in base model, curriculum and scoring metadata. Equal totals can conceal different probe outcomes. No cross-condition causal conclusion is asserted.
Archived harness verdicts
SZLHOLDINGS/a11oy-code-r1-candidate
Base: Qwen/Qwen2.5-Coder-3B-Instruct
3 / 6
3 of 6
Not a model accuracy estimate or deployment qualification.
The receipt does not establish the evaluated immutable model revision, full-response capture, matched decoding, independent repeated trials, or present runtime readiness.
Archived harness verdicts
Qwen/Qwen3-1.7B
Base: Qwen/Qwen3-1.7B
3 / 6
3 of 6
Not a model accuracy estimate or deployment qualification.
The receipt does not establish the evaluated immutable model revision, full-response capture, matched decoding, independent repeated trials, or present runtime readiness.
Archived harness verdicts
SZLHOLDINGS/code-master-v3-fresh
Base: Qwen/Qwen2.5-Coder-3B-Instruct
3 / 6
3 of 6
Not a model accuracy estimate or deployment qualification.
The receipt does not establish the evaluated immutable model revision, full-response capture, matched decoding, independent repeated trials, or present runtime readiness.
Archived harness verdicts
SZLHOLDINGS/doctrine-master-v1-candidate
Base: Qwen/Qwen3-1.7B
3 / 6
3 of 6
Not a model accuracy estimate or deployment qualification.
The receipt does not establish the evaluated immutable model revision, full-response capture, matched decoding, independent repeated trials, or present runtime readiness.
Archived harness verdicts
SZLHOLDINGS/hatun-24-doctrine-math-r1-candidate
Base: Qwen/Qwen3-1.7B
3 / 6
3 of 6
Not a model accuracy estimate or deployment qualification.
The receipt does not establish the evaluated immutable model revision, full-response capture, matched decoding, independent repeated trials, or present runtime readiness.
Archived harness verdicts
SZLHOLDINGS/measurement-master-v1-candidate
Base: Qwen/Qwen3-1.7B
3 / 6
3 of 6
Not a model accuracy estimate or deployment qualification.
The receipt does not establish the evaluated immutable model revision, full-response capture, matched decoding, independent repeated trials, or present runtime readiness.
Archived harness verdicts
SZLHOLDINGS/quant-master-4b-v1-candidate
Base: Qwen/Qwen3-4B
3 / 6
3 of 6
Not a model accuracy estimate or deployment qualification.
The receipt does not establish the evaluated immutable model revision, full-response capture, matched decoding, independent repeated trials, or present runtime readiness.
Archived harness verdicts
SZLHOLDINGS/quant-master-v1-candidate
Base: Qwen/Qwen3-1.7B
5 / 6
5 of 6
Not a model accuracy estimate or deployment qualification.
The receipt does not establish the evaluated immutable model revision, full-response capture, matched decoding, independent repeated trials, or present runtime readiness.
Archived harness verdicts
SZLHOLDINGS/reasoning-master-v1-candidate
Base: Qwen/Qwen3-1.7B
3 / 6
3 of 6
Not a model accuracy estimate or deployment qualification.
The receipt does not establish the evaluated immutable model revision, full-response capture, matched decoding, independent repeated trials, or present runtime readiness.
The receipt does not establish the evaluated immutable model revision, full-response capture, matched decoding, independent repeated trials, or present runtime readiness.
Archived harness verdicts
SZLHOLDINGS/abstention-v2-candidate/final
Base: Qwen/Qwen3-1.7B
3 / 6
3 of 6
Not a model accuracy estimate or deployment qualification.
The receipt does not establish the evaluated immutable model revision, full-response capture, matched decoding, independent repeated trials, or present runtime readiness.