# Zyphra's ZOE runs 3.1x faster per user than stock SGLang on AgentX

> Original title: ZOE generates 3.1x faster per user than stock SGLang on one 8xMI300X node using @SemiAnalysis_ AgentX (a replay of multi-turn coding agen...

- Source: Zyphra
- Published: 2026-11-05T19:51:15.000Z
- Bad Influence: https://badinfluence.news/items/pxc3mvmuop1itrvzw32syy0j9
- Original: https://x.com/ZyphraAI/status/2118430967589470463

## Summary

Zyphra says its ZOE generates 3.1x faster per user than stock SGLang on one 8xMI300X node in the SemiAnalysis AgentX replay of multi-turn coding agent sessions. At 64 concurrent sessions, ZOE reaches 52.8 tokens/s per user at p90 generation speed, versus 16.9 tokens/s for stock SGLang.

## Post

ZOE generates 3.1x faster per user than stock SGLang on one 8xMI300X node using @SemiAnalysis_ AgentX (a replay of multi-turn coding agent sessions).

At 64 concurrent sessions p90 generation speed:

ZOE = 52.8 tokens/s per user
Stock SGLang = 16.9 tokens/s per user

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWYtfoYaYAAat_M.jpg&mode=thumb&exp=1794096000&sig=2334c790b7190782)