SPY-Bench - Distractors - Conflict AtA: leaderboard
SPY-Bench - Distractors - Conflict AtA: Privacy instruction score (0-100). SPY-Bench arXiv 2611.13997v1: personalized visual-privacy compliance, scores 0-100. Native distractor/scenario/conflict and ICS/RtA/AtA/HMS columns remain separate. Full official source model labels bind the evaluated checkpoints. Official repository release links bind four fine-tuned Qwen2-VL-7B-Instruct SFT/DPO/NCA/NCA-P checkpoints; they are distinct released models, not run wrappers around the original model. Self-moderation is an inference configuration of the original. Table 7 first row is an external official result and is not recounted. API effort is unreported.
Metric: Privacy instruction score (0-100). Source: arxiv.org. Saturation forecast: Estimated already saturated. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2 VL 7B Instruct | 88.51 | #2569 |
| 2 | Phi-4 Multimodal Instruct | 84.81 | #3914 |
| 3 | Qwen 2.5 VL 7B Instruct | 82.88 | #1842 |
| 4 | Gemini 2.0 Flash | 82.86 | #702 |
| 5 | Llama 3.2 Vision 11B Instruct | 81.07 | #3876 |
| 6 | Pixtral-12B | 75.22 | #2493 |
| 7 | InternVL2.5-78B | 72.14 | #1018 |
| 8 | Qwen 2.5 VL 72B Instruct | 71.71 | #845 |
| 9 | Qwen 2.5 VL 32B Instruct | 70.92 | #1055 |
| 10 | Mistral Small 3.1 | 68.51 | #1699 |
| 11 | O4 Mini (2025-04-16) | 66.01 | #362 |
| 12 | GPT-4o (2024-11-20) | 59.24 | #864 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=spy-bench-distractors-conflict-ata · How It Works · Data refreshed daily, snapshot 2026-11-12.