# Ant Ling reports three optimizations boosting inference throughput

> Original title: Three optimizations made the difference:

- Source: Ant Ling
- Published: 2026-10-29T05:41:15.000Z
- Bad Influence: https://badinfluence.news/items/awx0y01h8zllbh2uewjq5ka52
- Original: https://x.com/AntLingAGI/status/2115680340866191773

## Summary

Ant Ling says three optimizations raised its throughput, with humming MXFP4 MoE kernels adding 9.3% and online FP8 LM-head quantization adding 21.7% cumulatively. It says NEXTN-based MTP drafts and verifies multiple tokens per round, reducing the number of decoding steps while the first two cut the cost per step.

## Post

Three optimizations made the difference:
1. Humming MXFP4 MoE kernels: +9.3%
2. Online FP8 LM-head quantization: +21.7% cumulatively
3. NEXTN-based MTP: drafting and verifying multiple tokens per round
The first two reduce the cost per step. MTP reduces the number of steps.

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHVxnjNqbsAAOwlk.png&mode=thumb&exp=1793491200&sig=6ee248617735d9b5)

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHVxnlmFasAAVtDC.png&mode=thumb&exp=1793491200&sig=cb5a09168f2422bd)

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHVxnnQqa0AAFME3.png&mode=thumb&exp=1793491200&sig=6ddb9af77ad2da4b)