The short version
Token-timing visualizer that makes LLM throughput feel tangible.
From the HackerLinks archive
Token-timing visualizer that makes LLM throughput feel tangible.
The short version
Token-timing visualizer that makes LLM throughput feel tangible.
Why it caught our attention
Useful for sanity-checking speed claims and local-model tradeoffs.
Where it surfaced on Hacker News
Editorial paraphrase
A commenter linked their own similar simulator and liked using real model output to render token timing.
Original threadHow fast is N tokens per second really?