EN
RU
ES
ZH
FR
DE
PT
AR
HI
Benchmark 2025: TypeScript vs Python, Go, Rust, and Java — who is faster?
2025 has become a turning point for the development world. TypeScript, long perceived as a "superset" for frontend, is increasingly invading backend and system programming. But how competitive is it in terms of speed? We analyzed the latest results from independent benchmarks (The Benchmarks Game and community tests) and are ready to share unexpected conclusions.
Methodology and Scenarios
Tests were conducted on modern hardware (AMD Ryzen 9 7950X, 64 GB RAM). Three key scenarios were measured:
- Pure Computation — recursive Fibonacci, sorting, matrix multiplication.
- I/O and Networking — file read/write, HTTP requests, JSON parsing.
- Real Microservice — REST API with a database (PostgreSQL), 1000 concurrent requests.
Results: Where TypeScript Wins and Where It Loses
Here is how the rankings for overall performance turned out (1 being the fastest):
- Rust — 1st place. Still the king of raw speed. It outperforms TypeScript by 3-5 times on computations.
- Go — 2nd place. Minimal gap from Rust in I/O, but 2 times faster than TypeScript on matrix operations.
- Java (GraalVM) — 3rd place. Surprisingly strong performance in 2025 thanks to AOT compilation. TypeScript lags by 30-40%.
- TypeScript (Bun runtime) — 4th place. A sensation! Bun 1.2 makes TypeScript 10 times faster than Python and almost catches up to Java in JSON scenarios.
- Python — 5th place. Unchanged — the slowest, but the most flexible.
Why is TypeScript Progressing So Much?
The secret lies in the runtimes. In 2025, Bun and Deno have practically pushed Node.js out of benchmarks. They use the V8 engine with JIT compilation and native bindings. For web applications, TypeScript on Bun outperforms Python (FastAPI) by 8-12 times, and Go by only 15-20%.
An Unexpected Verdict for Developers
If you need maximum performance, choose Rust or Go. But if development speed, ecosystem, and "one language for everything" (frontend + backend) are important, TypeScript 2025 is a reasonable compromise. It doesn't catch up to Rust, but it already leaves Python behind and is breathing down Java's neck.
At PythonLib, we recommend not ignoring TypeScript even for data science: libraries like TensorFlow.js and ONNX Runtime work surprisingly fast in 2025. And for startups, where time to market is more important than milliseconds, TypeScript is the best choice.
"In 2025, TypeScript is not just 'everything' for the web. It is a real competitor to Python in the backend and Go in microservices. But you will have to pay for computation speed with memory: Bun consumes 40% more RAM than Go." — notes the benchmark author.
What's Next?
Stay tuned for updates on PythonLib — we have prepared a separate performance test of TypeScript vs Python in machine learning. Spoiler: the results will surprise you.
