# prime-rl with Mega MoE trains Qwen3 MoE models faster than TorchTitan

> Original title: Comparing prime-rl with Mega MoE against TorchTitan:

- Source: Prime Intellect
- Published: 2026-10-29T17:41:36.000Z
- Bad Influence: https://badinfluence.news/items/ifq1qmjnxue3wfmqe47r83243
- Original: https://x.com/PrimeIntellect/status/2115861621440680237

## Summary

Prime Intellect says prime-rl with Mega MoE completes training steps 48% faster than TorchTitan on Qwen3-30B-A3B. On Qwen3-235B-A22B, the same comparison shows steps 55% faster.

## Post

Comparing prime-rl with Mega MoE against TorchTitan:

• Training steps on Qwen3-30B-A3B are 48% faster
• Training steps on Qwen3-235B-A22B are 55% faster

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHV0M7qjbkAAgUL0.png&mode=thumb&exp=1793491200&sig=621c2ed164590ab6)