D2VBench - Survival Security: leaderboard

Metric: Value-alignment score (0-100) on the survival-security scenarios; 10,000 everyday value-dilemma scenarios answered free-form; GLM-4.6 and Qwen3-Max map each response to the matching options and score its coverage of annotated interpretability points over five dimensions (0-100), the two judges averaged. Source: arxiv.org. Saturation forecast: Around June 2027. 8 models tracked.

Top models

#ModelScore
1GPT-5.166.68
2Gemini 3 Pro (Preview)63.45
3DeepSeek R1 052862.21
4GLM-4.662.09
5MiniMax-M261.16
6Claude Haiku 4.5 (20251001)60.01
7Kimi K2 090559.42
8Seed-1.657.34

Interactive version: theaggregate.ai/benchmark?slug=d2vbench-survival-security · How It Works · Data refreshed daily, snapshot 2026-09-29.