Test-time scaling improves small VLMs on multilingual visual MCQ

The study examines whether test-time scaling (TTS) transfers to small open vision-language models using the EXAMS-V benchmark, comparing methods like self-consistency and PRM-guided beam search on Qwen2.5-VL-7B-Instruct and Qwen3.5-4B.

• Parseability is the largest factor; adding a standard answer cue and guided repair step prevents chains from failing to commit to an answer letter.
• Increasing the per-chain token limit from 1k to 2k recovers 3.7 pp, while sampling more chains (8 to 16) adds only 0.15 pp.
• Elaborate methods contribute little once chains have room to finish; PRM-guided beam search trails plain self-consistency by 0.39 pp at over eight times the cost.
• Neither a training-free generative critic nor a trained multimodal PRM beats majority vote across both policies.

The best configuration reaches 84.1% on the held-out ImageCLEF 2026 test split, ranking first on the Visual MCQ leaderboard.