> Data revision: baf7c2f4307d3a989a8fb9363df8000fda3d9310
> Source files:
- [hypotheses/README.md](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/hypotheses/README.md)
- [hypotheses/f04-models-routing.md](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/hypotheses/f04-models-routing.md)

# H40

**Family:** [F04](/families/f04/) — Models, efforts and routing

**Falsifiable proposition:** The newer Luna model has a better cost-success profile than the earlier Luna baseline on the same tasks and harness.

**Current evidence status:** NOT ESTABLISHED; model preference alone is not a measured comparison.

This proposition and status are transcribed from the source hypothesis register; any numeric target in the proposition is not a measured result. The register does not provide a separate falsification rule or next-test field for this entry. The linked family and round sources retain their original limits.

### Related hypotheses

- [H40](/hypotheses/h40/)

### Related experiments

- [x-models](/rounds/x-models/)

### Related claims

- [CL-WA-x-models-reported-x-models-adjacent-panel](/claims/cl-wa-x-models-reported-x-models-adjacent-panel/)
- [CL-WA-x-models-reported-x-models-astra](/claims/cl-wa-x-models-reported-x-models-astra/)
- [CL-WA-x-models-reported-x-models-overall](/claims/cl-wa-x-models-reported-x-models-overall/)
- [CL-WA-x-models-reported-x-models-primary-screen](/claims/cl-wa-x-models-reported-x-models-primary-screen/)
