VIA-Bench (LLM Judge) - Color Illusions: leaderboard
Metric: Accuracy (%) on the colour illusions questions, answer letter read from the response by a GPT-4.1-mini judge; VIA-Bench's 1,004 human-verified multiple-choice questions on illusory and anomalous images, every option set including a never-correct 'Not Sure' option; zero-shot with an answer-format instruction, temperature 0.8 where supported (model defaults otherwise), mean of five runs; higher is better. Source: arxiv.org. Saturation forecast: Not forecast. 24 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 56.82 | #64 |
| 2 | O3 | 51.74 | #121 |
| 3 | GPT-5 Mini | 49.25 | #176 |
| 4 | GLM-4.5V | 47.56 | #339 |
| 5 | Qwen 3 VL 30B A3B (Thinking) | 39.9 | #338 (Qwen 3 VL 30B A3B) |
| 6 | Qwen 3 VL 235B A22B (Thinking) | 36.42 | #228 (Qwen 3 VL 235B A22B) |
| 7 | Claude Opus 4.1 (20250805) | 31.24 | #126 |
| 8 | GPT-4o (2024-11-20) | 29.55 | #369 |
| 9 | Qwen 2.5 VL 72B Instruct | 29.15 | #364 |
| 10 | Claude 3.5 Sonnet (20241022) | 28.96 | #286 |
| 11 | Gemini 2.5 Pro | 28.46 | #145 |
| 12 | O4 Mini | 28.16 | #172 |
| 13 | Claude Sonnet 4 (20250514) | 22.59 | #211 |
| 14 | Qwen 2.5 VL 7B Instruct | 22.09 | #643 |
| 15 | GPT-4o ChatGPT | 12.24 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=via-bench-llm-judge-color-illusions · How It Works · Data refreshed daily, snapshot 2026-10-11.