mlsanityData to 14 Oct 2026

AutoRubric: Rubric-Based Generative Rewards for Faithful Multimodal Reasoning

arXiv 2510.14738 ↗First posted 16 Oct 2025cs.CLcited by 11

On the map: LLM reasoning → Multimodal reasoning

also in 1 more
Abstract (arXiv)

Multimodal large language models (MLLMs) have rapidly advanced from perception tasks to complex multi-step reasoning, yet reinforcement learning with verifiable rewards (RLVR) often leads to spurious reasoning since only the final-answer correctness is rewarded. To address this limitation, we propose AutoRubric, a framework that integrates RLVR with process-level supervision through automatically collected rubric-based generative rewards. Our key innovation lies in a scalable self-aggregation method that distills consistent reasoning checkpoints from successful trajectories, enabling problem-specific rubric construction without human annotation or stronger teacher models. By jointly leveraging rubric-based and outcome rewards, AutoRubric achieves state-of-the-art performance on six multimodal reasoning benchmarks and substantially improves reasoning faithfulness in dedicated evaluations.

Key facts. Cited by 11 papers in the map (data through 14 Oct 2026). Placed in Multimodal reasoning, one of 17 directions in LLM reasoning. Builds on 27 of its 41 references that are in the map. First posted on arXiv 16 Oct 2025.

Builds on

27 of 41 references.

Closest ancestors · 15

Same direction, newest first

  1. 2025Answer-Consistent Chain-of-thought Reinforcement Learning For Multi-modal Large Langauge Models11 Oct
  2. 2025GRPO-CARE: Consistency-Aware Reinforcement Learning for Multimodal Reasoning19 Jun
  3. 2025Perception-R1: Advancing Multimodal Reasoning Capabilities of MLLMs via Visual Perception Reward8 Jun
  4. 2025SRPO: Enhancing Multimodal LLM Reasoning via Reflection-Aware Reinforcement Learning2 Jun
  5. 2025Mixed-R1: Unified Reward Perspective For Reasoning Capability in Multimodal Large Language Models30 May
  6. 2025Advancing Multimodal Reasoning via Reinforcement Learning with Cold Start28 May
  7. 2025NoisyRollout: Reinforcing Visual Reasoning with Data Augmentation17 Apr
  8. 2025VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning10 Apr
  9. 2025SoTA with Less: MCTS-Guided Sample Selection for Data-Efficient Visual Reasoning Self-Improvement10 Apr
  10. 2025Skywork R1V: Pioneering Multimodal Reasoning with Chain-of-Thought8 Apr
  11. 2025OThink-MR1: Stimulating multimodal generalized reasoning capabilities via dynamic reinforcement learning20 Mar
  12. 2025VisualPRM: An Effective Process Reward Model for Multimodal Reasoning13 Mar
3 more
  1. 2025MM-Eureka: Exploring the Frontiers of Multimodal Reasoning with Rule-based Reinforcement Learning10 Mar
  2. 2025Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models9 Mar
  3. 2025Unlocking Multimodal Mathematical Reasoning via Process Reward Model8 Jan

Other references · 12

  1. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  2. 2025Qwen2.5-VL Technical Reportcited by 5,587
  3. 2024Large Language Monkeys: Scaling Inference Compute with Repeated Samplingcited by 991
  4. 2025Rubrics as Rewards: Reinforcement Learning Beyond Verifiable Domainscited by 366
  5. 2025R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimizationcited by 301
  6. 2024Mulberry: Empowering MLLM with o1-like Reasoning and Reflection via Collective Monte Carlo Tree Searchcited by 175
  7. 2025OpenRubrics: Towards Scalable Synthetic Rubric Generation for Reward Modeling and LLM Alignmentcited by 132
  8. 2025Checklists Are Better Than Reward Models For Aligning Language Modelscited by 128
  9. 2025Reinforcement Learning with Rubric Anchorscited by 113
  10. 2025VerIF: Verification Engineering for Reinforcement Learning in Instruction Followingcited by 44
  11. 2023A Baseline Analysis of Reward Models' Ability To Accurately Analyze Foundation Models Under Distribution Shiftcited by 18
  12. 2025Robust Answers, Fragile Logic: Probing the Decoupling Hypothesis in LLM Reasoningcited by 11

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 11.

11 papers.

  1. 2026Reproducing, Analyzing, and Detecting Reward Hacking in Rubric-Based Reinforcement Learningcited by 16
  2. 2026Not Every Rubric Teaches Equally: Policy-Aware Rubric Rewards for RLVRcited by 11
  3. 2026EBPO: Empirical Bayes Shrinkage for Stabilizing Group-Relative Policy Optimizationcited by 5
  4. 2026RuCL: Stratified Rubric-Based Curriculum Learning for Multimodal Large Language Model Reasoningcited by 5
  5. 2026From Holistic Evaluation to Structured Criteria: Rubrics Across the Evolving LLM Landscapecited by 4
  6. 2026DeltaRubric: Generative Multimodal Reward Modeling via Joint Planning and Verificationcited by 3
  7. 2026AnyAudio-Judge: A Dynamic Rubric-Based Benchmark and Evaluator for Audio Instruction Followingcited by 3
  8. 2026Reinforcement Learning with Evolving Rubrics as Rewards for Audio Reasoningcited by 3
  9. 2026Prioritizing the Best: Incentivizing Reliable Multimodal Reasoning by Rewarding Beyond Answer Correctnesscited by 2
  10. 2026R^2-Mem: Reflective Experience for Memory Searchcited by 2
  11. 2026RLVR2: Reinforcement Learning with Verifiable Rubric-based Rankingcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026StructReward: Efficient Structured Process Rewards for Self-Correcting Multimodal Reasoning
  2. 2025StructVRM: Aligning Multimodal Reasoning with Structured and Verifiable Reward Models
  3. 2025ARM-Thinker: Reinforcing Multimodal Generative Reward Models with Agentic Tool Use and Visual Reasoning
  4. 2025Self-Rewarded Multimodal Coherent Reasoning Across Diverse Visual Domains
  5. 2025Enhancing the Outcome Reward-based RL Training of MLLMs with Self-Consistency Sampling
  6. 2026SSL-R1: Self-Supervised Visual Reinforcement Post-Training for Multimodal Large Language Models
  7. 2026Seeing with You: Perception-Reasoning Coevolution for Multimodal Reasoning
  8. 2025GM-PRM: A Generative Multimodal Process Reward Model for Multimodal Mathematical Reasoning