# Prime Intellect's Mega MoE speeds up MoE layers up to 2.47×

> Original title: Comparing the time per MoE layer between Prime Mega MoE and our prime-rl baseline:

- Source: Prime Intellect
- Published: 2026-10-29T17:41:38.000Z
- Bad Influence: https://badinfluence.news/items/no1dy6h63i5blzxmvi4gf6zzk
- Original: https://x.com/PrimeIntellect/status/2115861631624434072

## Summary

Prime Intellect says its Prime Mega MoE runs the forward pass up to 3.11× faster and the backward pass up to 1.87× faster than its prime-rl baseline per MoE layer. With activation checkpointing, the full layer is up to 2.47× faster.

## Post

Comparing the time per MoE layer between Prime Mega MoE and our prime-rl baseline:

• The forward pass is up to 3.11× faster
• The backward pass is up to 1.87× faster
• The full layer with activation checkpointing is up to 2.47× faster

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHV0M8N2bsAA55hb.jpg&mode=thumb&exp=1793491200&sig=5f36250ed5f4f28e)