AIPulse

About AIPulse

AIPulse publishes independent, reproducible AI benchmark results and technical analysis. No vendor partnerships. No affiliate deals. Just data.

Mission

The AI space moves fast and hype moves faster. We benchmark local and cloud models across real-world tasks — reasoning, coding, hallucination resistance, long context — and publish the full methodology and scripts so anyone can reproduce the results.

Test Hardware

Machine
Apple Mac Mini (M4 Pro)
Chip
Apple M4 Pro
Memory
24GB unified (LPDDR5X)
Memory Bandwidth
273 GB/s
LLM Runtime
Ollama 0.9+
Cloud APIs
OpenAI, Google AI Studio

Open Source

All benchmark scripts are published on GitHub. If you find an error or want to contribute a new test, open a pull request.

github.com/venkatlab/aipulse