See-or-Guess VQA - Multi-Signal: leaderboard

Metric: Multi-Signal variant (source question fused with sub-questions elicited turn by turn in a chat): 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 October 2027. 11 models tracked.

Top models

#ModelScore
1Qwen 3.5 9B58.7
2Gemini 2.5 Pro58.5
3Qwen 3.5 397B A17B54.8
4Claude Sonnet 4.654.3
5GPT-5 Mini49.4
6Gemini 3.1 Flash Lite43.7
7InternVL3.5-8B36.1
8Phi-4 Multimodal Instruct29.1

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