# vLLM V0 to V1: Correctness Before Corrections in RL

DevFeed: [vLLM V0 to V1: Correctness Before Corrections in RL](<https://devfeed.tech/articles/vllm-v0-to-v1-correctness-before-corrections-in-rl-7047.md>)

Original publisher: [Read original article](<https://huggingface.co/blog/ServiceNow-AI/correctness-before-corrections>)

Author: Rafael Pardinas; Ehsan Kamalloo

Published: 2026-05-06T19:06:55Z

Content type: article

Language: en

Sources: [Hugging Face - Blog](<https://devfeed.tech/sources/hugging-face-blog.md>)

Topics: [vllm](<https://devfeed.tech/topics/vllm.md>), [Back end](<https://devfeed.tech/topics/backend.md>), [Inference](<https://devfeed.tech/topics/inference.md>), [benchmarking](<https://devfeed.tech/topics/benchmarking.md>), [Optimization](<https://devfeed.tech/topics/optimization.md>)

Tags: [backend](<https://devfeed.tech/tags/backend.md>), [comparison](<https://devfeed.tech/tags/comparison.md>), [inference](<https://devfeed.tech/tags/inference.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [rl](<https://devfeed.tech/tags/rl.md>), [vllm](<https://devfeed.tech/tags/vllm.md>)

## AI overview

This article explains how a vLLM V1 migration was brought to parity with a vLLM V0 reference for reinforcement-learning training. The authors fixed processed rollout logprobs, V1 runtime defaults, the inflight weight-update path, and final-projection precision before changing the RL objective.

## Source excerpt

TL;DR. vLLM V1 matched our vLLM V0 reference after we fixed four things: processed rollout logprobs, V1-specific runtime defaults, the inflight weight-update path, and the fp32 used for the final projection. We fixed the backend behavior before changing the RL objective. The reference run used vLLM ; the V1 runs used vLLM . Figure 1 shows the final result. The red run is the initial V1 attempt, and the green run is the final V1 run after the fixes described below.