# prime‑rl 0.10.0 adds new IPO loss and speed boosts

> Prime Intellect released prime‑rl v0.10.0 and verifiers v0.4.0, featuring a faster default RL loss (IPO) and up to 15% training speed gains.

Oossa · 2026-10-19 · https://oossa.com/en/prime-rl-0-10-0-adds-new-ipo-loss-and-speed-boosts

Prime Intellect announced prime‑rl version 0.10.0 together with verifiers 0.4.0. The update makes the new IPO (INTELLECT Policy Optimization) loss the default for asynchronous reinforcement learning, replacing the older DPPO‑based loss. It also includes performance tweaks that make training up to 1.15× faster on large models like Qwen3‑235B‑A22B, and adds support for new architectures such as DeepSeek V4 Flash and Qwen‑VL.

## The facts

- prime‑rl v0.10.0 and verifiers v0.4.0 released (date not given)
- Training speed improvement: 1.15× faster on Qwen3‑235B‑A22B compared to torchtitan

## Why it matters

Researchers can train large language models a bit quicker and with a more stable RL loss, reducing the chance of training collapse.

## Sources & references

1. [ResearchOCT 19TH, 2026prime-rl 0.10.0: faster training, our new RL loss IPO, and one unified eval stack](https://www.primeintellect.ai/blog/prime-rl-0-10-verifiers-0-4) – Prime Intellect

Last updated: 2026-10-19