mlsanityData to 14 Oct 2026

Beyond Multiple Choice: Verifiable OpenQA for Robust Vision-Language RFT

arXiv 2511.17405 ↗First posted 21 Nov 2025cs.CL cs.AIcited by 1

On the map: Multimodal LLMs / vision-language models → VLM benchmarks and visual reasoning

Abstract (arXiv)

Multiple-choice question answering (MCQA) has been a popular format for evaluating and reinforcement fine-tuning (RFT) of modern multimodal language models. Its constrained output format allows for simplified, deterministic automatic verification. However, we find that the options may leak exploitable signals, which makes the accuracy metrics unreliable for indicating real capabilities and encourages explicit or implicit answer guessing behaviors during RFT. We propose ReVeL (Rewrite and Verify by LLM), a framework that rewrites multiple-choice questions into open-form questions while keeping answers verifiable whenever possible. The framework categorizes questions according to different answer types, apply different rewriting and verification schemes, respectively. When applied for RFT, we converted 20k MCQA examples and use GRPO to finetune Qwen2.5-VL models. Models trained on ReVeL-OpenQA match MCQA accuracy on multiple-choice benchmarks and improve OpenQA accuracy by about six percentage points, indicating better data efficiency and more robust reward signals than MCQA-based training. When used for evaluation, ReVeL also reveals up to 20 percentage points of score inflation in MCQA benchmarks (relative to OpenQA), improves judging accuracy, and reduces both cost and latency. We will release code and data publicly.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in VLM benchmarks and visual reasoning, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 38 of its 45 references that are in the map. First posted on arXiv 21 Nov 2025.

Builds on

38 of 45 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025VisualSimpleQA: A Benchmark for Decoupled Evaluation of Large Vision-Language Models in Fact-Seeking Question Answering9 Mar
  2. 2025Automated Generation of Challenging Multiple-Choice Questions for Vision Language Model Evaluation6 Jan

Other references · 36

  1. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  2. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  3. 2025Qwen2.5-VL Technical Reportcited by 5,587
  4. 2018Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challengecited by 4,886
  5. 2023MMBench: Is Your Multi-modal Model an All-around Player?cited by 2,141
  6. 2023MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGIcited by 2,080
  7. 2022Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answeringcited by 1,919
  8. 2024MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmarkcited by 1,848
  9. 2024Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scalingcited by 1,599
  10. 2016A Diagram Is Worth A Dozen Imagescited by 804
  11. 2024MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmarkcited by 512
  12. 2023Large Language Models Are Not Robust Multiple Choice Selectorscited by 486
24 more
  1. 2021Inter-GPS: Interpretable Geometry Problem Solving with Formal Language and Symbolic Reasoningcited by 383
  2. 2024Measuring short-form factuality in large language modelscited by 373
  3. 2024MLLM-as-a-Judge: Assessing Multimodal LLM-as-a-Judge with Vision-Language Benchmarkcited by 372
  4. 2025R1-Onevision: Advancing Generalized Multimodal Reasoning through Cross-Modal Formalizationcited by 370
  5. 2025VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learningcited by 323
  6. 2024MME-RealWorld: Could Your Multimodal LLM Challenge High-Resolution Real-World Scenarios that are Difficult for Humans?cited by 292
  7. 2023Large Language Models Sensitivity to The Order of Options in Multiple-Choice Questionscited by 263
  8. 2024CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMscited by 234
  9. 2021GeoQA: A Geometric Question Answering Benchmark Towards Multimodal Numerical Reasoningcited by 228
  10. 2024When Benchmarks are Targets: Revealing the Sensitivity of Large Language Model Leaderboardscited by 137
  11. 2025Can MLLMs Reason in Multimodality? EMMA: An Enhanced MultiModal ReAsoning Benchmarkcited by 134
  12. 2024Open-LLM-Leaderboard: From Multi-choice to Open-style Questions for LLMs Evaluation, Benchmark, and Arenacited by 76
  13. 2024Artifacts or Abduction: How Do LLMs Answer Multiple-Choice Questions Without the Question?cited by 71
  14. 2025Which of These Best Describes Multiple Choice Evaluation with LLMs? A) Forced B) Flawed C) Fixable D) All of the Abovecited by 52
  15. 2025Right Answer, Wrong Score: Uncovering the Inconsistencies of LLM Evaluation in Multiple-Choice Question Answeringcited by 28
  16. 2023Assessing the Quality of Multiple-Choice Questions Using GPT-4 and Rule-Based Methodscited by 24
  17. 2024Beyond Probabilities: Unveiling the Misalignment in Evaluating Large Language Modelscited by 19
  18. 2024II-Bench: An Image Implication Understanding Benchmark for Multimodal Large Language Modelscited by 16
  19. 2025EgoThinker: Unveiling Egocentric Reasoning with Spatio-Temporal CoTcited by 16
  20. 2024Enhancing Distractor Generation for Multiple-Choice Questions with Retrieval Augmented Pretraining and Knowledge Graph Integrationcited by 11
  21. 2024Wait, that's not an option: LLMs Robustness with Incorrect Multiple-Choice Optionscited by 11
  22. 2025Reasoning Models are Test Exploiters: Rethinking Multiple-Choicecited by 11
  23. 2025SATA-BENCH: Select All That Apply Benchmark for Multiple Choice Questionscited by 9
  24. 2025None of the Above, Less of the Right: Parallel Patterns between Humans and LLMs on Multi-Choice Questions Answeringcited by 5

Built on by

Cited by year

Citing papers per year: 2025: 1.

1 paper.

  1. 2025Unexplored flaws in multiple-choice VQA make benchmarking unreliablecited by 1

Similar, not cited

Close in wording, no citation either way

  1. 2026FAU at ImageCLEF 2026 Task on Multimodal Reasoning Robust Candidate Scoring and Concise Multilingual Visual Answering
  2. 2025Benchmarking and Mitigating MCQA Selection Bias of Large Vision-Language Models
  3. 2024Addressing Blind Guessing: Calibration of Selection Bias in Multiple-Choice Question Answering by Video Language Models
  4. 2026VOILA: Value-of-Information Guided Fidelity Selection for Cost-Aware Multimodal Question Answering
  5. 2026Rethinking Multiple-Choice Questions for RLVR: Unlocking Potential via Distractor Design
  6. 2026MetaRA: Metamorphic Robustness Assessment for Multimodal Large Language Model-based Visual Question Answering Systems
  7. 2024Answer, Assemble, Ace: Understanding How LMs Answer Multiple Choice Questions
  8. 2024LLM Distillation for Efficient Few-Shot Multiple Choice Question Answering