Benchmarks

Performance comparisons between MediatorLite and MediatR across four scenarios: simple request dispatch, single pipeline behavior, multiple pipeline behaviors, and notification publishing.

Last updated: 2026-03-29 — automated via CI.

Environment

BenchmarkDotNet v0.15.8, Linux Ubuntu 24.04.4 LTS (Noble Numbat)
AMD EPYC 7763 2.45GHz, 1 CPU, 4 logical and 2 physical cores
.NET SDK 10.0.201
  [Host]     : .NET 10.0.5 (10.0.5, 10.0.526.15411), X64 RyuJIT x86-64-v3
  Job-YFEFPZ : .NET 10.0.5 (10.0.5, 10.0.526.15411), X64 RyuJIT x86-64-v3

IterationCount=10  WarmupCount=3

---

Simple Request

Baseline scenario: a single request dispatched to a single handler with no pipeline behaviors.

Method	Mean	Error	StdDev	Ratio	RatioSD	Gen0	Allocated	Alloc Ratio
MediatR_SimpleRequest	113.0 ns	1.91 ns	1.26 ns	1.00	0.02	0.0196	328 B	1.00
MediatorLite_SimpleRequest	140.4 ns	1.38 ns	0.91 ns	1.24	0.02	0.0153	256 B	0.78

---

Single Pipeline Behavior

Request dispatched through one open pipeline behavior (simulating a logging or tracing layer).

Method	Mean	Error	StdDev	Ratio	Gen0	Allocated	Alloc Ratio
MediatR_WithBehavior	205.5 ns	1.90 ns	1.26 ns	1.00	0.0381	640 B	1.00
MediatorLite_WithBehavior	280.3 ns	1.94 ns	1.15 ns	1.36	0.0348	584 B	0.91

---

Multiple Pipeline Behaviors

Request dispatched through three behaviors (logging, validation, metrics) — a typical production pipeline.

Method	Mean	Error	StdDev	Ratio	Gen0	Allocated	Alloc Ratio
MediatR_WithMultipleBehaviors	349.2 ns	3.36 ns	2.22 ns	1.00	0.0639	1072 B	1.00
MediatorLite_WithMultipleBehaviors	530.9 ns	3.61 ns	2.15 ns	1.52	0.0505	856 B	0.80

---

Notifications

A notification published to three handlers, comparing MediatR against MediatorLite’s Sequential and Parallel execution strategies.

Method	Mean	Error	StdDev	Ratio	Gen0	Allocated	Alloc Ratio
MediatR_Notification	190.1 ns	2.42 ns	1.60 ns	1.00	0.0367	616 B	1.00
MediatorLite_Sequential_Notification	253.0 ns	2.95 ns	1.76 ns	1.33	0.0134	224 B	0.36
MediatorLite_Parallel_Notification	312.9 ns	0.96 ns	0.63 ns	1.65	0.0176	296 B	0.48

---

Key Takeaways

MediatorLite consistently allocates less memory across every scenario. The notification sequential mode is especially notable — near-identical latency to MediatR but less than half the allocations. The ValueTask-based pipeline and source-generated dispatch path trade raw per-call speed for a predictable, low-allocation footprint.

---

Running Benchmarks Locally

cd tests/MediatorLite.Benchmarks
dotnet run --configuration Release -- --filter '*' --exporters json markdown --memory

Results are written to BenchmarkDotNet.Artifacts/results/.

REST API Benchmarks (Production-Like)

The repository also includes a REST API benchmark project at tests/MediatorLite.RestApiBenchmarks. This suite compares MediatorLite and MediatR in end-to-end API scenarios backed by SQLite:

• In-process transport (TestServer) and real localhost transport (Kestrel)
• Read-heavy and write-heavy API operations
• Concurrency scenarios with multiple in-flight requests

Run read/write benchmarks:

dotnet run -c Release --project tests/MediatorLite.RestApiBenchmarks -- --filter '*RestApiReadWriteBenchmarks*'

Run concurrency benchmarks:

dotnet run -c Release --project tests/MediatorLite.RestApiBenchmarks -- --filter '*RestApiConcurrencyBenchmarks*'