Steps to reproduce:

# Open the attached user's Project “LeakMemory”
# Select Android build target
# Add the default scene in the build settings window
# Check "Development Build"
# Check "Autoconnect Profiler"
# Build the application
# Run the application and let it idle
# Using the "Memory Profiler" tool take screenshots every 5 minutes
# Observe and compare the memory consumption

Expected results: Memory consumption doesn’t exponentially increase within minutes
Actual results: Memory consumption increased by 40 to 80 megabytes in 15 minutes after letting the application idle

Reproducible with: 2021.3.16f1, 2022.1.24f1, 2022.2.2f1, 2023.1.0a25
Couldn't test with: 2020.3.40f1 due to memory profiling not connecting to devices

Reproduces on these devices:
VLNQA00013 - Galaxy S6 Edge+ (SM-G928F), CPU: Exynos 7 Octa (7420), GPU: Mali-T760, OS: 7.0.0
VLNQA00317 - Vivo NEX 3 5G (V1924A), CPU: Snapdragon 855+, GPU: Adreno 640, OS: 9.0.0
N/A, Google Pixel 3 (Pixel 3), Android 12, CPU: Snapdragon 845, GPU: Adreno (TM) 630
VLNQA00413 - Samsung Galaxy Note10+ (SM-N975F), OS: 9, CPU: Exynos 9 Series 9825, GPU: Mali-G76
VLNQA00372 - Galaxy S21 5G (SM-G991U), CPU: Snapdragon 888, GPU: Adreno 660, OS: 11
VLNQA00202 - Samsung Galaxy Note 9 USA (SM-N9600), CPU: Snapdragon 845 SDM845, GPU: Adreno 630, OS: 9.0.0
VLNQA00177 - Xiaomi Mi Note 2 (Mi Note 2), OS: 6.0.1, CPU: Snapdragon 820 MSM8996, GPU: Adreno 530
VLNQA00231 - HUAWEI Mate 20 Pro (LYA-L29), OS: 9.0.0, CPU: HiSilicon Kirin 980, GPU: Mali-G76
VLNQA00406 - Galaxy Z Flip3 5G (SM-F711B), OS: 11, CPU: Snapdragon 888, GPU: Adreno 660
N/A, Samsung Galaxy S9 (SM-G960F), Android 8.0.0, CPU: Exynos 9 Series 9810, GPU: Mali-G72
VLNQA00358 - iPhone 12, 14.1 iOS

Reproducible on: macOS 12.4 (Intel)

Notes: Reproducible with IOS metal; Android OpenGLES3, Vulkan, IL2CPP, Mono, release mode with keystore