SPY-Bench - Distractors - Nonconflict ICS: leaderboard
SPY-Bench - Distractors - Nonconflict ICS: 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 | GPT-4o (2024-11-20) | 32.6 | #864 |
| 2 | O4 Mini (2025-04-16) | 25.87 | #362 |
| 3 | Qwen 2.5 VL 32B Instruct | 21.73 | #1055 |
| 4 | Mistral Small 3.1 | 20.9 | #1699 |
| 5 | Qwen 2.5 VL 72B Instruct | 20.01 | #845 |
| 6 | Pixtral-12B | 16.15 | #2493 |
| 7 | InternVL2.5-78B | 14.63 | #1018 |
| 8 | Gemini 2.0 Flash | 11.54 | #702 |
| 9 | Llama 3.2 Vision 11B Instruct | 11.4 | #3876 |
| 10 | Qwen 2.5 VL 7B Instruct | 5.51 | #1842 |
| 11 | Phi-4 Multimodal Instruct | 4.96 | #3914 |
| 12 | Qwen 2 VL 7B Instruct | 3.94 | #2569 |
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-nonconflict-ics · How It Works · Data refreshed daily, snapshot 2026-11-12.