See-or-Guess VQA - Vision-Grounded: leaderboard

Metric: Vision-Grounded variant (questions Gemini-2.5-Pro writes from the image alone, no source text): accuracy (%) with the image over 540 images from 21 public benchmarks in six reasoning categories; Qwen3-14B judges answer equivalence (yes/no), lenient extraction of answers after a missing </think> tag. Source: arxiv.org. Saturation forecast: Around April 2028. 11 models tracked.

Top models

#ModelScore
1Qwen 3.5 9B42
2Qwen 3.5 397B A17B38.4
3Gemini 2.5 Pro35.9
4GPT-5 Mini33.3
5Claude Sonnet 4.631.6
6Gemini 3.1 Flash Lite27.3
7InternVL3.5-8B15.9
8Phi-4 Multimodal Instruct10

Interactive version: theaggregate.ai/benchmark?slug=see-or-guess-vqa-vision-grounded · How It Works · Data refreshed daily, snapshot 2026-09-29.