> Data revision: baf7c2f4307d3a989a8fb9363df8000fda3d9310
> Source files:
- [hypotheses/f11-claim-reconciliation.md](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/hypotheses/f11-claim-reconciliation.md)
- [hypotheses/README.md](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/hypotheses/README.md)

# F11 — Claimed cross-cutting results

> Family source text is narrative from the repository. Any numeric value in that text remains source-reported unless an exact claim ledger entry supplies it.

# F11 — Claimed cross-cutting results

**Status: INTERIM.** This family collects historical summaries that span experiments; a summary is not an additional experiment. The claims below stay separate from the round records they may describe. The L0 reconciliation adds reproducible analysis from published sanitized CSVs for selected R72, R74, R69, and R57 populations. That does not imply every row is independently joined to the canonical official outcome export; each row below preserves its remaining join and scope limits.

## Hypotheses

- **H111 — Staged and direct model comparison.** On a shared task set, a staged Kogen route may help a weaker builder solve failures while direct Sol high may have a higher pass rate, lower cost per pass, or longer wall time. Falsify with exact task-matched outcomes, cost, and wall receipts for every arm.
- **H112 — Route cost and quality frontier.** Across a shared task set, direct Sol, route-v3, and direct Luna max may occupy different pass-rate and cost-per-pass points. Falsify with one reconciled cohort, compatible cost accounting, and task-level outcomes.
- **H113 — R72 same-model comparison.** On R72 P1, the Kogen Luna-max ladder may outperform direct Codex Luna max. Falsify with a complete cell join and a prespecified paired analysis.
- **H114 — R70 language choice.** The R70 implementation-stack comparison and selected RvE sample may identify a better language for Kogen. Evaluate with a parity-controlled comparison whose task, prompt, harness, and repair loop are matched; a selected sample alone cannot determine a broad language winner.
- **H115 — Cache-hit explanation.** A Kogen cache-hit rate below Codex's may explain much of the token-spend difference. Falsify with request-level records from matched epochs and compatible cache/token definitions, then relate cache changes to cost.
- **H116 — Cache-change effect.** A cache change may increase cache use. Falsify with a defined metric and baseline, matched request traffic, and before/after records.

## Claim and run register

The reproduction fields below distinguish the original design from available public exports. `Not recovered` means the public record does not establish the value. A round page is the authority for its run history; claims link to `results/claim-ledger.jsonl`.

| experiment_id / claim | Arm; tasks; model and effort | Revision; host and epoch | planned / started / finished / graded / ITT n | Outcome and status | Correction or supersession | Claim and receipt |
|---|---|---|---|---|---|---|
| Historical staged/direct summary (H111) | Arm mapping, task IDs, model/effort, and comparison direction not recoverable. | Harness and Kogen pins, host, and epoch not identified. | Not recoverable. | Historical summary reports 85% and 89% direct-model pass rates, about 2.5× lower cost per pass, and 1.5–2× wall time for an unspecified direction. **Source-reported, not reproducible from public data.** No comparison is supportable. | Keep distinct from R53, R64, and later held-out rounds; no exact cohort was recovered. | [`CL-H111-historical-staged-and-direct-comparison`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl) (**INTERIM**); no public round receipt. |
| Historical route-v3 cost frontier (H112) | Direct Sol, route-v3, direct Luna max; 18 shared tasks are reported, but membership, model receipts, and effective effort are not recovered. | Harness/Kogen pins, host, and epoch not identified. | Shared-task count is source-reported; started, finished, graded, and ITT counts not recoverable. | Historical summary reports 76%/$0.38 per pass for direct Sol, 70%/$0.38 for route-v3, and 50%/$0.047 for Luna max, with four Rails feature tasks unsolved. **Source-reported, not reproducible from public data.** This does not establish a current frontier. | The cohort and cost inputs do not join to public exact-cell records; do not treat this as a current frontier estimate. | [`CL-H112-historical-route-cost-frontier`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl) (**INTERIM**); no public round receipt. |
| R72 P1 and partial P3 (H113, H136, H138) | P1: Kogen Luna-max ladder vs direct Codex Luna max. P3: mixed Kogen ladder vs direct Codex Sol high. The L0 CSVs publish exact recovered task and cell IDs for these subsets. | Kogen [`260a73bc06be16e563cb4e56a168ffd078cec677`](https://github.com/KogenAI/kogen-ex/commit/260a73bc06be16e563cb4e56a168ffd078cec677); comparator harness pin and worker/epoch assignments not recovered. | L0 reproduces P1 n=84 per arm and partial P3 n=13/11. Initial plan 456, complete start/finish totals, and exact ITT partition remain source-reported or unrecovered. The L0 rows do not match the canonical public outcome export. | P1: 64/84 vs 60/84; P3: 12/13 vs 10/11. **Reproducible from L0 CSV; not independently joined to the official outcome export; INTERIM, descriptive only; no superiority claim.** | Early stop and superseded Kogen revision; keep P1/P2/P3 separate and do not pool with R74. | [`CL-WA-r72-lifecycle-and-outcome`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl); [R72 page](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r72/README.md); [L0 results and CSVs](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/l0-reconcile/README.md). |
| R69 L0 reconciliation | Four arms: SPEC, APPROACH, ACCEPT, NONE; exact task-level rows and outcomes are published in the L0 CSV. | No separate Kogen source commit recorded; historical run pins remain incomplete. | L0 reproduces the 400 assigned rows and 396-row included ITT, including 327 internal grades and 69 ITT failures. The refreshed 2026-10-07 public export exact-joins 150/327 by `cell_id`. | The arm counts and selected intervals reproduce from the CSV. **INTERIM; partial official-outcome join; not a verified comparison.** | Preserve the source cohort and keep the partial public grade join visible. | [`CL-WA-r69-source-reported-denominator-variants`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl); [R69 page](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r69/README.md); [L0 results and CSVs](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/l0-reconcile/README.md). |
| R57 pooled L0 reconciliation | Recovered pooled cohort has 311 exact rows across 11 tasks and two arms. | Source-declared strata and selected estimator are retained; one planned desktop Mac slot has no recovered cell ID or grade. | L0 reproduces the 311 recovered rows and reports matching public grades for those rows; the full planned 312-cell denominator is incomplete by one slot. | Shell-only 114/155 and default tools 111/156; published interim bound reproduces from the CSV. **INTERIM; scoped recovered population; no broad tools-effect claim.** | Keep component rounds and the later replacement cohort separate. | [`CL-WA-r57-source-reported-denominator-variants`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl); [R57 page](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r57/README.md); [L0 results and CSVs](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/l0-reconcile/README.md). |
| R70 original RvE (H114) | Direct Codex `gpt-6-luna`, max; four implementation stacks on tasks 1, 3, 4, and 6. | Kogen commit not applicable. No Codex harness Git commit; the public record has a fingerprint for only a subset. Host is not a pooled timing unit. | 40 planned/identified/officially graded; five additional ungraded deliveries are **source-reported, not reproducible from public data**, and lack joinable IDs, so ITT is unresolved. | Full passes: Elixir 2/10, Go 7/10, Rust 8/10, TypeScript/Bun 7/10. **CONFOUNDED; descriptive only.** Task/skeleton parity limits the stack comparison. | Keep the original cohort as-graded; the fixed-skeleton rerun is a separate cohort. | [`CL-WA-r70-rve-r70-original-rve-elixir-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl), [`CL-WA-r70-rve-r70-original-rve-go-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl), [`CL-WA-r70-rve-r70-original-rve-rust-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl), [`CL-WA-r70-rve-r70-original-rve-ts-bun-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl); [R70 RvE page](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r70-rve/README.md); [test-count ledger](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/test-counts.jsonl). |
| R70 reported selected ten-pick sample (H114) | The selection is described as five random and five strategic picks; exact tasks, model/effort, and scored cohort are not present in the public R70 page. | Revision, harness, host, and epoch are not recovered for the selected-pick statement. | Start, finish, grade, and ITT counts are not recovered. The ten-pick description is **source-reported, not reproducible from public data**. | No result for this selected sample is available to support language choice. | Keep the selected-pick description separate from original, rerun, extension, task-8, and compile cohorts. | No claim ID assigned; [R70 RvE page](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r70-rve/README.md), [test-count ledger](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/test-counts.jsonl). |
| R70 fixed-skeleton rerun (H114) | Direct Codex `gpt-6-luna`, max; task 1 Elixir and task 4 across four stacks. | Kogen commit not applicable. Per-cell harness version not verified. EU and US hosts are split by stack; no pooled timing comparison. | 15 official rerun grades; exact start/finish/ITT partition is not reported as a single reconciled count. | Full passes: Elixir 3/6, Go 3/3, Rust 3/3, TypeScript/Bun 2/3; 11/15 overall. **CONFOUNDED; descriptive sensitivity check.** Admission controls failed for three task-4 stacks. | Do not replace or pool with the original cohort. | [`CL-WA-r70-rve-rerun-r70-rve-rerun-elixir-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl), [`CL-WA-r70-rve-rerun-r70-rve-rerun-go-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl), [`CL-WA-r70-rve-rerun-r70-rve-rerun-rust-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl), [`CL-WA-r70-rve-rerun-r70-rve-rerun-ts-bun-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl); [rerun page](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r70-rve-rerun/README.md); [test-count ledger](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/test-counts.jsonl). |
| R70 task 2/5/7 extension (H114) | Direct Codex `gpt-6-luna`, max; Rust, Elixir, Go, and TypeScript/Bun. | Kogen commit not applicable. Per-cell harness version unavailable. EU for Rust/Elixir; US for Go/TypeScript-Bun. | 24 preregistered cells plus five post-hoc additions; 29 official grades. | Registered reps 31–32: Rust 5/6, Elixir 4/6, Go 4/6, TypeScript/Bun 6/6. Including post-hoc rep 33: Rust 6/7, Elixir 5/8, Go 6/8, TypeScript/Bun 6/6. **VALID, descriptive; small samples and selected task templates.** | Keep registered reps, post-hoc rep 33, and the original/RvE rerun as distinct cohorts. | [`CL-WA-r70-rve-ext-r70-rve-ext-elixir-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl), [`CL-WA-r70-rve-ext-r70-rve-ext-go-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl), [`CL-WA-r70-rve-ext-r70-rve-ext-rust-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl), [`CL-WA-r70-rve-ext-r70-rve-ext-ts-bun-full-pass`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl); [extension results](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r70-rve-ext/RESULTS.md); [test-count ledger](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/test-counts.jsonl). |
| Sol-medium language replication (H91, H114) | Direct Codex `gpt-6.1-sol`, medium; Rust, Go, and TS-Bun across six task identities. | Kogen not applicable; new Rust cells ran on EU and Go/TS-Bun on US; compare wall time within host only. | 27 new cells; 31 reused Sol cells retained as a separate prior cohort in the 58-cell combined summary. | Raw totals: Rust 18/20, Go 18/20, TS-Bun 15/18. Post-hoc equal-task means: 90.3%, 88.9%, and 83.3%. | **VALID round, DESCRIPTIVE**; equal-task formula is post-hoc and the combined four-pass branch remains under-specified. | [`CL-SOL-LANG-REPLICATION-POSTHOC-EQUAL-TASK`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl) (**STATUS_ONLY**); [round page](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/lang-sol-replication/README.md); [F08 language evidence](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/hypotheses/f08-language-spec.md). |
| R70 task-8 v2 smokes (H114) | Rust and Go smoke attempts; direct Codex `gpt-6-luna`, max. No scored language comparison. | Kogen commit not applicable; these are direct Codex task builds. No exact per-cell Codex harness Git commit or CLI version is recorded. | Maximum planned 4; started/finished/officially graded smoke cells 2; scored cells and scored ITT n=0. | Both Rust and Go smoke cells failed the release gate at 1/25 tests. **DESCRIPTIVE, unscored smoke only.** | Task 8 is excluded from language inference. | [`CL-WA-r70-task8v2-smoke-codex-gpt-6-luna-max-default-r70-8-go-v2-r1`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl), [`CL-WA-r70-task8v2-smoke-codex-gpt-6-luna-max-default-r70-8-rust-v2-r1`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl); [task-8 v2 page](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r70-rve-task8v2/README.md); [test-count ledger](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/test-counts.jsonl). |
| R74 design and withdrawal (H115, H138) | Registered Kogen planner Sol high with Luna max/Sol low/medium/high builders; direct Codex matched builder settings. Twelve-task panel. | Kogen [`cde7a380455e9793cb1fc8315791bc9beffb863b`](https://github.com/KogenAI/kogen-ex/commit/cde7a380455e9793cb1fc8315791bc9beffb863b); exact harness commit and per-cell use not public. | L0 reproduces the recovered 167 internal scored rows, 9 smoke attempts, and coverage counts from published CSVs. The public tagged outcome export remains empty; 231 retained slots lack recovered IDs and terminal statuses. | **WITHDRAWN.** The recovered subset is a documentary reconciliation, not an efficacy result or full retained-cohort reconstruction. | Preserve amendment and withdrawal history. Do not substitute the recovered subset for the full cohort or pool it with R72/frontier. | [`CL-WA-r74-source-reported-denominator-variants`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl); [`CL-WA-r74-public-export-coverage`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl); [R74 page](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r74/README.md); [L0 results and CSVs](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/l0-reconcile/README.md). |
| Historical cache-hit comparison (H115) | Kogen vs Codex; epochs, request populations, models, effort, and task set are not recovered. | Harness/Kogen pins and hosts not identified. | Not recovered. | Historical summaries report 26.5% Kogen, 93% Codex, and about 70% for a separate local aggregate. **Source-reported, not reproducible from public data.** Token-spend explanation is unverified. | Different or missing epochs and definitions prevent direct comparison with T98. | [`CL-H115-historical-cache-hit-comparison`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl) (**INTERIM**); no public raw request receipt. |
| T98 cache smoke (H115) | One R74-style diagnostic cell; Sol medium is reported; exact task and per-request model/effort are not published. | Kogen [`3644d35d7b83f171cf4461f931885daccad1014a`](https://github.com/KogenAI/kogen-ex/commit/3644d35d7b83f171cf4461f931885daccad1014a); exact harness commit not supplied; EU smoke. | One smoke cell; 20 develop requests, 18 with usage. Scored/ITT n=0. Request counts and rates are **source-reported, not reproducible from public data**. | Weighted cache hit 67.0%; turn-3+ minimum 0%; below the 95% gate. Telemetry only, not an efficacy result. | Separate from historical 26.5%/93% and the local ~70% summary. | [`CL-WA-t98-cache-smoke-reported-observation`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl) (**INTERIM**); [T98 page](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/t98-cache-smoke/README.md). |
| T100 cache lane (H116) | No executed model/effort, task, or arm recorded. | No execution revision or host. | Planned n not supplied; started, finished, graded, and ITT n=0 in the public status record. | **NOT-RUN.** No cache-change result. | The historical 19% statement has no metric definition, baseline, period, or request denominator. | [`CL-WA-t100-cache-lane-reported-observation`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl); [T100 page](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/t100-cache-lane/README.md). |

## Design implications

- Keep each historical claim tied to its original cohort and distinguish a Kogen recipe from a direct-model baseline.
- Keep R70 original, rerun, extension, task-eight smoke, and compile-time measurements as separate endpoints and cohorts.
- Keep cache telemetry separate from Build efficacy, and require matched epochs and compatible token accounting before explaining cost differences.

## Explicit non-implications

- The source-reported R72 difference does not establish superiority or a current-version effect.
- The R70 cohorts do not establish a broad language ranking or select a Kogen implementation language.
- T98's cache smoke does not show that a cache change increases success or lowers cost; T100 provides no evidence because it did not run.
- None of the historical cross-cutting percentages adds new experimental observations beyond its underlying run or claim record.

## Smallest open tests

- Reconcile the historical H111/H112 claims to immutable cell IDs, official outcomes, task pairs, full pins, model/effort receipts, and compatible cost/wall inputs; otherwise retire the numeric comparisons.
- For a language decision, register a matched implementation cohort with identical task behavior, prompt, agent loop, tools, and repair budget; include exact harness pins and a complete ITT join.
- For cache claims, run a fixed request workload before and after the change and publish sanitized request-level token/cache records with the metric definition and denominator.


## Related structured records

### Related hypotheses

- [H111](/hypotheses/h111/)
- [H112](/hypotheses/h112/)
- [H113](/hypotheses/h113/)
- [H114](/hypotheses/h114/)
- [H115](/hypotheses/h115/)
- [H116](/hypotheses/h116/)

### Related experiments

- [l0-reconcile](/rounds/l0-reconcile/)
- [lang-sol-replication](/rounds/lang-sol-replication/)
- [r70-rve-task8v2](/rounds/r70-rve-task8v2/)
- [r72](/rounds/r72/)

### Related claims

- [CL-H111-historical-staged-and-direct-comparison](/claims/cl-h111-historical-staged-and-direct-comparison/)
- [CL-H112-historical-route-cost-frontier](/claims/cl-h112-historical-route-cost-frontier/)
- [CL-H115-historical-cache-hit-comparison](/claims/cl-h115-historical-cache-hit-comparison/)
- [CL-H116-historical-cache-change](/claims/cl-h116-historical-cache-change/)
- [CL-SOL-LANG-REPLICATION-POSTHOC-EQUAL-TASK](/claims/cl-sol-lang-replication-posthoc-equal-task/)
- [CL-WA-r57-source-reported-denominator-variants](/claims/cl-wa-r57-source-reported-denominator-variants/)
- [CL-WA-r69-source-reported-denominator-variants](/claims/cl-wa-r69-source-reported-denominator-variants/)
- [CL-WA-r70-rve-ext-r70-rve-ext-elixir-full-pass](/claims/cl-wa-r70-rve-ext-r70-rve-ext-elixir-full-pass/)
- [CL-WA-r70-rve-ext-r70-rve-ext-go-full-pass](/claims/cl-wa-r70-rve-ext-r70-rve-ext-go-full-pass/)
- [CL-WA-r70-rve-ext-r70-rve-ext-rust-full-pass](/claims/cl-wa-r70-rve-ext-r70-rve-ext-rust-full-pass/)
- [CL-WA-r70-rve-ext-r70-rve-ext-ts-bun-full-pass](/claims/cl-wa-r70-rve-ext-r70-rve-ext-ts-bun-full-pass/)
- [CL-WA-r70-rve-r70-original-rve-elixir-full-pass](/claims/cl-wa-r70-rve-r70-original-rve-elixir-full-pass/)
- [CL-WA-r70-rve-r70-original-rve-go-full-pass](/claims/cl-wa-r70-rve-r70-original-rve-go-full-pass/)
- [CL-WA-r70-rve-r70-original-rve-rust-full-pass](/claims/cl-wa-r70-rve-r70-original-rve-rust-full-pass/)
- [CL-WA-r70-rve-r70-original-rve-ts-bun-full-pass](/claims/cl-wa-r70-rve-r70-original-rve-ts-bun-full-pass/)
- [CL-WA-r70-rve-rerun-r70-rve-rerun-elixir-full-pass](/claims/cl-wa-r70-rve-rerun-r70-rve-rerun-elixir-full-pass/)
- [CL-WA-r70-rve-rerun-r70-rve-rerun-go-full-pass](/claims/cl-wa-r70-rve-rerun-r70-rve-rerun-go-full-pass/)
- [CL-WA-r70-rve-rerun-r70-rve-rerun-rust-full-pass](/claims/cl-wa-r70-rve-rerun-r70-rve-rerun-rust-full-pass/)
- [CL-WA-r70-rve-rerun-r70-rve-rerun-ts-bun-full-pass](/claims/cl-wa-r70-rve-rerun-r70-rve-rerun-ts-bun-full-pass/)
- [CL-WA-r70-rve-ungraded-delivery-itt-gap](/claims/cl-wa-r70-rve-ungraded-delivery-itt-gap/)
- [CL-WA-r70-task8v1-smoke-codex-gpt-6-luna-max-default-r70-8-go-r90](/claims/cl-wa-r70-task8v1-smoke-codex-gpt-6-luna-max-default-r70-8-go-r90/)
- [CL-WA-r70-task8v1-smoke-codex-gpt-6-luna-max-default-r70-8-rust-r90](/claims/cl-wa-r70-task8v1-smoke-codex-gpt-6-luna-max-default-r70-8-rust-r90/)
- [CL-WA-r70-task8v2-smoke-codex-gpt-6-luna-max-default-r70-8-go-v2-r1](/claims/cl-wa-r70-task8v2-smoke-codex-gpt-6-luna-max-default-r70-8-go-v2-r1/)
- [CL-WA-r70-task8v2-smoke-codex-gpt-6-luna-max-default-r70-8-rust-v2-r1](/claims/cl-wa-r70-task8v2-smoke-codex-gpt-6-luna-max-default-r70-8-rust-v2-r1/)
- [CL-WA-r72-lifecycle-and-outcome](/claims/cl-wa-r72-lifecycle-and-outcome/)
- [CL-WA-r74-source-reported-denominator-variants](/claims/cl-wa-r74-source-reported-denominator-variants/)
- [CL-WA-t100-cache-lane-reported-observation](/claims/cl-wa-t100-cache-lane-reported-observation/)
- [CL-WA-t98-cache-smoke-reported-observation](/claims/cl-wa-t98-cache-smoke-reported-observation/)
