AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Omni-RRM: Advancing Omni Reward Modeling via Automatic Rubric-Grounded Preference Synthesis

Zicheng Kong, Dehua Ma, Zhenbo Xu, Anwen Yang, Yiwei Ru, Haoran Wang, Zixuan Zhou, Fuqing Bie, Liuyu Xiang, Huijia Wu, Jian Zhao, Zhaofeng He

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Unified any-modality language models · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add GRPO after SFT (ablation)Omni-RRM-7B · Overall (5 benchmarks)preference accuracy (%): 62.6 -> 70.4Improved
add GRPO after SFT (ablation)Omni-RRM-7B · MM-RewardBenchpreference accuracy (%): 61.0 -> 72.9Improved
add GRPO after SFT (ablation)Omni-RRM-7B · ShareGPT-Videopreference accuracy (%): 70.5 -> 80.2Improved
add rubric rationales during SFT+RL (ablation)Omni-RRM-7B · Overall (4-benchmark ablation suite)preference accuracy (%): 64.4 -> 71.8Improved
add rubric reward term in GRPO (ablation)Omni-RRM-7B · Overall (4-benchmark ablation suite)preference accuracy (%): 68.1 -> 71.8Improved
add rubric rationales during SFT+RL (ablation)Omni-RRM-3B · Overall (4-benchmark ablation suite)preference accuracy (%): 58.9 -> 65.0Improved
add rubric reward term in GRPO (ablation)Omni-RRM-3B · Overall (4-benchmark ablation suite)preference accuracy (%): 62.9 -> 65.0Improved
train Omni-RRM on omni-modal data instead of audio only (ablation)Omni-RRM-3B · Audio-HH-RLHFpreference accuracy (%): 63.8 -> 65.1Improved
use Omni-RRM-7B as Best-of-N selector instead of self-consistencyQwen2.5-Omni-7B · MMMUaccuracy (%): 54.5 -> 56.4Improved
use Omni-RRM-7B as Best-of-N selector instead of self-consistencyQwen2.5-Omni-7B · Video-MMEaccuracy (%): 52.4 -> 54.1Improved
use Omni-RRM-7B as Best-of-N selector instead of self-consistencyQwen2.5-Omni-7B · AVQAaccuracy (%): 77.7 -> 80.3Improved
apply Omni-RRM-7B post-training to Qwen2.5-Omni-7B backboneQwen2.5-Omni-7B · Overall (5 benchmarks)preference accuracy (%): 60.2 -> 70.4Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy