Uni-SafeBench - Image-Unsafe VQA: leaderboard
Uni-SafeBench (2026) tests the safety of unified multimodal models, vision-language models and LLMs on malicious requests across understanding and generation tasks, with a judge that checks both the context of each response and its intrinsic content. This board is the Both-Safe rate in percent on visual questions with a malicious image and a benign text; higher is better.
Metric: Both-Safe rate (%) on the 467 image-unsafe visual question answering queries of Uni-SafeBench (malicious image, benign text), judged by the Uni-Judger framework (malicious-intent extractor plus safety detector on Qwen3-VL-8B-Instruct), a response counts as safe only when it is safe or a refusal in the context of the extracted intent and its own content is not intrinsically unsafe, mean over repeated judge runs; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 14 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Chameleon (Uni-SafeBench checkpoint unspecified) | 98.9 | |
| 2 | GPT-4o | 96.1 | #751 |
| 3 | Gemini 2.5 Flash Image | 92.1 | |
| 4 | Qwen 2.5 VL 7B Instruct | 89.7 | #1902 |
| 5 | DeepSeek-VL-7B-Chat | 83.3 | #3756 |
| 6 | Tar-7B | 67.9 | |
| 7 | Show-o2-7B | 60.4 | |
| 8 | Harmon-1.5B | 60.2 | |
| 9 | Tar-1.5B | 43.9 | |
| 10 | BAGEL-7B-MoT | 40 | #4619 |
| 11 | DeepSeek-VL-7B-Base | 33.4 | |
| 12 | Janus-Pro-7B | 32.3 | #6857 |
| 13 | Emu3 | 22.1 | |
| 14 | Janus-flow | 19.3 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=uni-safebench-image-unsafe-vqa · How It Works · Data refreshed daily, snapshot 2026-11-21.