DMC-CF - Double Intervention (L2): leaderboard

DMC-CF (2026) double-intervention counterfactuals: questions generated from causal graphs of real-world videos that add a second hypothetical change, which either breaks the outcome or leaves it intact. Yes/no accuracy in percent; higher is better.

Metric: Accuracy (%) on the L2 set: questions adding a second intervention generated from an extracted causal graph that either breaks the outcome (answer no) or leaves it intact (answer yes), yes/no counterfactual questions about 1,614 real-world causal event videos (videos given natively to Gemini and Qwen3-VL, as at most 16 frames to GPT and Claude); higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 4 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1DMC-CF Claude-4 (checkpoint unspecified)58.14
2GPT-5.255.2#190
3Qwen 3 VL 235B A22B (Thinking)54.93#462
4Qwen 3 VL 235B A22B Instruct48.44#541

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=dmc-cf-double-intervention-l2 · How It Works · Data refreshed daily, snapshot 2026-11-22.