Foundgine
End-to-end evidence · Run 4 · 18 August 2026

Measure graph execution at scale: can semantic traversal stay fast?

Run 4 measures the customer-exposure workflow across 10, 100, 1,000 and 10,000 customers at concurrency 8, 16, 32 and 64, comparing GraphQL + Hot Chocolate + EF Core with MCP + Foundgine.

Inspect the matrix evidence →All benchmark runs
Why this run existsRun 4 isolates graph execution performance after the earlier agent-work experiments and asks whether semantic traversal scales across larger fixtures and concurrent workloads.
What was comparedThe same customer-exposure-review workflow through GraphQL + Hot Chocolate + EF Core and MCP + Foundgine, across the full 4 × 4 workload/concurrency matrix.
What it meansRun 4 is the graph-execution scalability result: the semantic path sustains materially higher throughput in the published matrix while preserving the agent-facing workflow.
Published finding
What the evidence says

Across the published 4 × 4 matrix, Foundgine sustained materially higher agent-flow throughput than the conventional path. At 10 customers/C8, the published RPS was about 1,488 versus 302 (4.94×). The aggregate contains 32 rows covering all 16 cells and both implementations, using only the agent samples rather than the separate single-request baseline samples.

Key measurements
toolCallsGraphQL agent vs MCP agent
contextAgent-only matrix
success32 aggregate rows
example10/C8: ~302 → ~1,488 RPS
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