Round r58x-studio
R58x Studio Elixir ladder lane
Source-reported narrativeRound README and results text are source material. Numeric values that are not tied to a claim ledger record remain source-reported.
Round source README · rounds/r58x-studio/README.md
Round date: HISTORICAL (before 2026-10-09) Publication badge: DESCRIPTIVE / INCOMPLETE (PILOT); KEPT FOR AUDIT Why not VALID: The historical inventory retains PILOT because of DESIGN_CHANGE, EVIDENCE_PARTIAL, NO_PREREG. See the round evidence and limitations below. Recomputation status: OBSERVED SOURCE ONLY
Recomputation update: The mined raw records now provide per-cell outcome, manifest token-counter, and wall-time recomputation for the observed cohort. This does not reconstruct missing planned cells, a full intention-to-treat denominator, or model execution.
Status
PILOT
- NO_PREREG — A dated pre-registration is not established in the public record.
- DESIGN_CHANGE — The source update supersedes the early blocked checkpoint.
- EVIDENCE_PARTIAL — The mined source now reconciles 36 scored cells and two smokes with per-cell grades; full request history and planned/ITT receipts remain unavailable.
Required reproduction metadata
- Kogen commit: `d8f29faab1a28a3b52c2cc70ec308cbd9f360450`, the source pin recorded for the Studio lane.
- Harness commit: not recorded in the cited round source as a Git commit; any adapter or runtime fingerprints are not Git commits.
- Task IDs:
elx-07-cli-stats,elx-12-retry-api-deprecation,elx-port-erase-account,elx-port-board-publish-unpublish-public-boundary,elx-02-ingest-supervision, andelx-04-queue-backpressure. - Raw records: safe per-cell records and joined official grades are in data/mined/r58x-studio.jsonl.gz; per-arm outcomes, token totals, wall medians, and comparisons are in recomputed.json. Request logs and transcripts are not included.
Terminology: public round glossary.
STATUS: PILOT
Data completeness: N/A (no captured public run-record rows are assigned to this round ID; source-reported activity is not evidence of zero historical execution). Measurement contract; declared public-record gaps.
Pre-registered: not established in the public record.
Related hypotheses: H77 (cross-stack task coverage), H87 (round design/status), H88 (intention-to-treat handling), and H138 (matched model/effort baselines).
Date: 6 October 2026 (source-reported).
Question: How did a Kogen ladder compare descriptively with direct Codex Sol high on the selected hard Elixir tasks?
Design: Source-reported design/status, not reproducible from public data. The source summary describes six hard Elixir tasks, a Kogen ladder, and direct Codex Sol high with three repetitions per task and arm. A later source update supersedes the early blocked checkpoint.
Historical source report: two sandbox smokes were reported. An early checkpoint showed zero of 36 scored cells released; a later update reports all 36 scored rows finished, 18 per arm, by the duplicate trim, with the smoke cells separate. The cell-level official outcome join is now in the recomputed section below.
Lifecycle counts
Historical source-reported lifecycle: Planned scored cells = 36. An early checkpoint reported released = 0; the later source update supersedes that checkpoint and reports 36/36 scored rows finished, 18 per arm. The mined store now has 38 official grades (36 scored and two smokes); the ITT partition remains unreconciled.
Reproduction
- Exact Kogen commit: Kogen pin
d8f29faaresolves to `d8f29faab1a28a3b52c2cc70ec308cbd9f360450`. The separate Codex CLI source or binary revision is not recovered. - Exact Kogen harness commit: see the full source pin above. For a separate comparator harness, its exact source or binary revision is not recovered.
- Model and effort: Kogen ladder versus direct Codex Sol high. Exact Kogen per-stage model/effort receipts and effective effort are unavailable.
- Task IDs:
elx-07-cli-stats,elx-12-retry-api-deprecation,elx-port-erase-account,elx-port-board-publish-unpublish-public-boundary,elx-02-ingest-supervision, andelx-04-queue-backpressure. - Exact command: The exact dispatch command is not recovered. This page is not independently rerunnable from the public snapshot.
- Raw record location: Archived round packet
rounds/r58x/and Studio result familyr58x/studio/; the mined safe records, joined official grades, usage/wall fields, and patch hashes are now in data/mined/r58x-studio.jsonl.gz; full request and all-attempt records are not included. - Reproduction procedure: To reproduce the status, the mined records now identify 36 scored cells and two smokes and join their official outcomes by immutable cell ID. Full request/all-attempt history, complete per-stage model/effort receipts, and the launch command remain unavailable.
Limits
The earlier zero-release snapshot is superseded as to execution completion. Exact-ID outcomes are recomputed below; this pilot does not support a winner, cost, or speed conclusion.
No measured superiority claim is made from this pilot. The outcomes are now recomputed for the observed 36-cell scored cohort; planned/ITT coverage and full execution receipts remain unresolved.
The current public run-record export contains no entries assigned to r58x-studio; that export does not establish that no historical run occurred.
Recomputed from raw records
`recomputed.json`, mined cell records, and the token audit back observed official outcomes, per-arm pass counts/rates, manifest token counters, and wall medians. The source contains 38 rows (fail 7, pass 31); overall pass rate is 81.6% (31/38) among pass/fail grades. Per-arm detail and coverage counts are in the JSON.
No shared cell IDs are present in the committed public grade export; the source-only cohort is checked against the page's reported figures below. No shared public-export cell IDs are available for a paired outcome comparison. No complete published metadata/manifest token pairs are available for this round (0 missing manifest usage; 38 missing published metadata). Missing counters remain unknown, not zero. This round also has 38 raw-only or non-public rows; they are kept separate from the published-export denominator. Historical cohort check: the mined data contain 36 scored cells and 2 smoke cells. This matches the published 36+2 split; scored arm counts: codex-sol-high 17/18, kogen-best-d8f 12/18. Separate smoke cell IDs:
codex__gpt-6.1-sol__high__default__elx-07-cli-stats__r1-codex-sol-high-r58x-studio-smokehc3-ladder__gpt-6-luna__max__default__elx-07-cli-stats__r1-kogen-best-d8f-r58x-studio-smoke
Captured run records
Standard-schema records assigned to this round in the indexed snapshot.
No per-round file is listed. Check the run-record index for the complete partition list.
Reproduction inventory
Published analysis replay and full experiment replay are reported separately.
- Replay type
unavailable- Script
- No per-round analysis script is published.
- Inputs
- No identified public input bundle.
- Full execution replay
- Unavailable
- Scope
- No public per-round analysis script with a complete, identified input bundle is published. See the round README, MEASURED.md, and MISSING.md for recorded scope and gaps; historical execution inputs/toolchains/grading are not bundled.
View the complete per-round reproduction inventory · Repository reproduction guide