Foundgine
End-to-end evidence · Run 3 · 17 August 2026

Measure efficiency: what happens when the workflow itself becomes larger?

Run 3 deliberately increases the amount of agent/tool work represented by the scenario. It asks whether the semantic path can reduce the amount of context and tool coordination without introducing a proportionally larger execution penalty.

Inspect the matrix evidence →All five runs
Why this run existsIf the value of semantic execution is partly about reducing orchestration work, the effect should become visible when the workflow has more steps to coordinate. This run isolates that efficiency question from the later graph-performance benchmark.
What was comparedThe same customer-exposure-review business process through conventional and semantic agent flows.
What it meansRun 3 is where the efficiency story becomes more compelling: the amount of agent-side work falls materially without the large latency penalty seen in Run 1.
Published finding
What the evidence says

The conventional flow used 56 tool calls versus 32 for Foundgine (42.9% fewer), and estimated context load fell from 7,568 to 4,364 tokens (42.3% lower). Average wall-clock time was almost flat: 14.64 ms versus 15.16 ms (+3.6%). Both flows reached the same final state.

Key measurements
toolCalls56 → 32
context7,568 → 4,364 est. tokens
wall14.64 → 15.16 ms
sameStateTrue
How to use this run

Read the finding first, then inspect the exact evidence.

This page tells the story of the experiment. The benchmark explorer contains the interactive workload/concurrency matrix and selected-cell comparison.

The benchmark distinguishes measured timings and tool calls from its estimated context-load heuristic.

← Benchmark landing page · How it works · Architecture