After purchasing a new VPS, most users' first reaction is the same: is this machine worth the price? The configuration parameters advertised by the vendor look good, but how is the actual performance? Is it over-sold? How fast is the network to China? These questions cannot be answered by simply looking at the configuration list. At this point, running several performance testing scripts becomes essential.

However, there are simply too many testing scripts on the market. YABS, Fusion, LemonBench, benchmark.sh, SuperBench… each tool has its own focus, and beginners often don't know which one to choose, while experienced users may need to switch back and forth between multiple scripts. This article will explain the most mainstream VPS performance testing scripts clearly, helping you find the one that best suits your needs.

YABS: The Internationally Recognized Comprehensive Benchmark

YABS stands for "Yet Another Benchmark Script." Although the name is casual, it is currently the most popular one-stop benchmarking script in the global VPS community. This script is maintained by the developer masonr and has a high level of activity on GitHub.

YABS' popularity stems from its comprehensive testing dimensions and standardized results. A single command performs three tasks sequentially: testing CPU single-core and multi-core performance using Geekbench 6, testing disk read and write speeds at different block sizes using fio, and testing network upload and download speeds to multiple nodes globally using iperf3. A complete run takes approximately 10 to 30 minutes.

YABS' CPU benchmark scores are highly valuable. Geekbench 6 is a cross-platform, authoritative benchmark tool, and YABS directly calls its command-line version, allowing for comparison with scores from other VPS users worldwide. Generally, a single-core score below 400 indicates weak CPU performance, 400 to 800 is average, 800 to 1200 is good, and over 1200 is excellent. For disk performance, YABS tests read and write speeds at 4K, 64K, 512K, and 1M block sizes, allowing users to determine whether the hard drive is an HDD, a regular SSD, or NVMe.

Running YABS is very simple:

curl -sL yabs.sh |

If the network is down, you can use an alternative address:

wget -qO- https://raw.githubusercontent.com/masonr/yet-another-bench-script/master/yabs.sh |

Who is YABS suitable for? If you need a standardized benchmark report that allows you to compare your VPS performance with global VPS users, or if you are making a VPS purchase decision and need to quantitatively compare the performance of machines from different providers, YABS is the most reliable choice. It does not depend on any third-party libraries and can run directly on almost all Linux distributions.

Fusion Monster: A "Full-Stack Evaluation Toolbox" Combining the Strengths of Many

If YABS is characterized by "standardization," then Fusion Monster is characterized by "comprehensiveness."

Fusion Monster (ECS, short for VPS Fusion Monster Server Test Script) is an integrated, comprehensive performance evaluation script created and open-sourced by developer spiritLHLS. Its core concept is "combining the strengths of many"—integrating various independent testing tools scattered throughout the open-source community into a unified entry point. Users don't need to search for and run multiple scripts individually; a single operation can complete a full-stack evaluation from hardware benchmarks to network quality.

The testing dimensions covered by Fusion Monster are quite impressive: system information detection (OS, kernel, CPU model, memory, virtualization type, disk, IP/ASN, etc.), CPU benchmark testing (single-core/multi-core computing performance), disk I/O testing, three-network speed testing, streaming media unlock detection, backhaul route tracing, and IP quality assessment—a total of seven dimensions.

Fusion Monster's most practical feature is its three-network backhaul route testing. It can automatically detect the backhaul path from the VPS to the backbone networks of China Telecom, China Unicom, and China Mobile, and identify the line type. This is crucial for businesses targeting domestic users—whether it uses CN2 GIA, CN2 9929, or the ordinary 163 line directly determines the latency and stability of access for domestic users.

Fusion Monster offers two running modes: sequential execution (average about 7 minutes) and parallel execution (average about 5 minutes, only suitable for high-performance machines). The main menu adopts an interactive design with seven functional areas, supporting three levels of nested selection.

Command to run the Affinity Monster Shell version:

curl -L https://gitlab.com/spiritysdx/za/-/raw/main/ecs.sh -o ecs.sh && chmod +x ecs.sh && ecs.sh

It's worth noting that the Affinity Monster project is currently migrating entirely from Shell to Go. The Go version eliminates references to third-party shell files, supports running with non-root privileges, and minimizes environment changes. The Shell version is currently in maintenance mode and no new features are being developed. New users are recommended to use the Go version, available at https://github.com/oneclickvirt/ecs.

Who is Affinity Monster suitable for? If you need a comprehensive report covering all dimensions of data, including hardware, network, and IP quality, or if your VPS is primarily for domestic users and you need to focus on understanding the quality of routers on the three major Chinese internet networks (China Unicom, China Telecom, and China Unicom), Affinity Monster is a highly efficient choice. A single run provides almost all key data, eliminating the hassle of switching between multiple scripts.

LemonBench: A Powerful Network Testing Tool for Domestic Users

LemonBench (also known as LBench or LemonBench) is a performance testing tool designed for Linux servers. Unlike the comprehensive offerings of YABS and FusionBench, LemonBench has a more focused approach—its specialty is testing domestic network environments.

LemonBench's core feature is its three-network backhaul route detection. The script uses traceroute to probe the path of data packets returning from the VPS to the backbone networks of China Telecom, China Unicom, and China Mobile, and automatically identifies the line type based on the AS number. For example, 4809 usually indicates CN2 GIA, 9929 indicates CN2 secondary, and 4134 indicates a regular 163 connection. This automatic identification capability allows users to quickly determine network quality without manually analyzing routing tables.

The command to run LemonBench for a quick test:

curl -L https://raw.githubusercontent.com/LemonBench/LemonBench/main/LemonBench.sh | -s fast

However, it's worth noting that LemonBench is currently maintained infrequently, and the last major update to the GitHub repository was quite early. Compatibility issues may occur when running it on newer systems. It's recommended to use its results as a reference indicator and perform secondary verification using tools like MTR.

Who is LemonBench suitable for? If you need to quickly screen the network quality of your VPS and determine whether it uses optimized routes like CN2, LemonBench remains a convenient tool. However, if you're looking for more comprehensive test coverage or a more frequently updated tool, FusionBench is a better alternative.

bench.sh: A classic, lightweight introductory choice

bench.sh is a VPS testing script written by teddysun and has been active in the VPS community for many years. Its positioning is "compact, basic, and one-click"—the script is small, doesn't require a lot of dependencies, and clearly lists the core data after execution.

The `bench.sh` test consists of three main parts: system information (CPU model, number of cores, whether AES-NI is supported, total disk space, memory, Swap, virtualization type, BBR status, etc.), disk I/O speed (averaged after three sequential writes using the `dd` command), and network speed test (upload and download speeds to multiple nodes globally).

Running `bench.sh`:

wget -qO- bench.sh |

The main difference between `bench.sh` and YABS is that `bench.sh` does not test CPU performance (Geekbench) or perform return route tracing. It is more lightweight and faster, but its test dimensions are relatively basic. Fortunately, the version number is still being updated; we can see a 2026 version number, indicating that it is still being maintained.

Who is `bench.sh` suitable for? For novice users first encountering VPS, running `bench.sh` to quickly understand the machine's basic information is the lowest-cost way to get started. Experienced users also frequently use it to quickly check the configuration of new machines.

How to choose?

Back to the initial question: Which should you choose: YABS, Amalgam, or LemonBench? Actually, there's no need to get hung up on choosing between these two or three tools. Each has its strengths, and using them in combination is the most efficient approach:

- Quick Start: Run `bench.sh` first to understand your system configuration and basic disk I/O in two or three minutes.

- Standardized Benchmarking: Run YABS to get authoritative CPU scores from Geekbench 6 and globally comparable network speed test data.

- Domestic Network Testing: Use FusionBench or LemonBench to test the backhaul routes of the three major Chinese networks (ISPs) to assess line quality. FusionBench is updated more frequently and has more comprehensive coverage, making it the recommended choice.

- Comprehensive Report: Run FusionBench directly (Go version recommended) to get all the data on hardware performance, IP quality, network stability, etc., all at once.

If you want to further improve testing efficiency, you can learn about NodeQuality—it combines the core functions of YABS, IPQuality, NetQuality, and FusionBench. One run, one report, and seamless testing make it ideal for users who need to frequently evaluate VPS. Jtti's official website help center has detailed NodeQuality usage tutorials for reference.

Regardless of which tool you choose, it's recommended to preview any scripts from the internet before running them to ensure their reliability. After all, security should always be the top priority.