Viral Thread Explains LLM Inference Metrics: TTFT, TPS and TPOT

A popular thread series by @abhijithneil breaks down key LLM inference metrics: TTFT dominated by the parallel prefill phase, and TPOT (inter-token latency) governing streaming generation speed, e.g. 20ms per token equals 50 tokens per second.

2026-09-04 ~ 2026-09-04 · 4 related posts

• LLM decoding explained: prefill reads in parallel, decode writes token by token — abhijithneil · 2026-09-04
• LLM inference metrics explained: what TTFT, TPS, TPOT actually measure — abhijithneil · 2026-09-04
• Streaming LLM metrics explained: TPOT is inter-token latency, the real streaming speed — abhijithneil · 2026-09-04
• TPOT of 20ms means 50 tokens/sec: a thread untangling LLM inference speed metrics — abhijithneil · 2026-09-04