AetherGate context benchmark

Five fictional workloads, two product arms, three replicates. We publish outcomes including failures. No savings or accuracy claim is made until the complete matrix has been measured and graded.

Baseline uses one model; Foreman uses a two-model workflow. Costs are AetherGate catalog estimates from reported usage, not provider invoices. Synthetic operational review notes expand the original authored corpus. Provider automatic caching and unsupported temperature controls are limitations.

• Northwind Logistics — 60,688 characters; 30.8% topic relevance. Smoke blocked: baseline measured; Foreman specialist execution failed. Full 120-run matrix not launched. Human grading pending.
• Marrow Health — 60,732 characters; 15.4% topic relevance. Not run: full matrix gated on successful two-arm smoke.
• Tessellate Studio — 63,155 characters; 45.5% topic relevance. Not run: full matrix gated on successful two-arm smoke.
• Bellwether Reporting — 77,707 characters; 21.7% topic relevance. Not run: full matrix gated on successful two-arm smoke.
• Open Arbor — 59,886 characters; 55.6% topic relevance. Not run: full matrix gated on successful two-arm smoke.