quantmodel — Benchmark evaluator hosts and choose private-pilot provider

Objective

Benchmark evaluator hosts and choose private-pilot provider.

Notes: Operator selected the existing free Render service. Live evidence: 5 controlled suspend/resume cold samples (p50 36.994s, p95 48.559s), 30 authenticated warm samples (p50 5.204s, p95 5.617s), zero failures; free plan, exact-origin CORS, disabled signup, missing/malformed auth rejection, and two-user RLS CRUD/cross-owner denial verified. Portable pg_dump restore remains t295 because this runner cannot reach Supabase PostgreSQL ports.

Approach

Reuses existing report research/findings/evaluator-host-benchmark-t294.md — tables/metrics preserved verbatim via source_report pointer. Body below summarizes grounded evidence from that report and adjacent artifacts.

Files / code / data changed

No commit SHA confidently associated with t294 in sampled log [reconstructed — uncertain].
Source report: research/findings/evaluator-host-benchmark-t294.md
Task-history entry: research/findings/task-history.json#t294

Results

Conclusions and metrics are in research/findings/evaluator-host-benchmark-t294.md; see detail page for preserved tables/metrics.

Conclusions

[reconstructed — uncertain] Outcome inferred from status done and task note.

Problems / follow-ups

See source report or PR discussion for follow-ups. No unsupported follow-ups fabricated.