CDR-Bench - Order-Agnostic Recipes: leaderboard

Metric: Recipe success RS@3 (%) on the order-agnostic mapper-recipe tasks of CDR-Bench (3,462 data-refinement tasks over web, LaTeX, RAG-preparation and privacy text with 29 Data-Juicer operators and deterministic reference outputs): a task counts when the model's refined text or KEEP/DROP judgment exactly matches the reference within 3 attempts; non-thinking mode, baseline prompt; higher is better. Source: arxiv.org. Saturation forecast: Around 2032. 11 models tracked.

Top models

#ModelScore
1Gemma 4 31B (IT) (Non-reasoning)25.55
2GLM-5 (Non-reasoning)24.8
3Kimi K2.6 (Non-reasoning)24.14
4DeepSeek V4 Flash (Non-reasoning)23.94
5Qwen 3.6 Max23.94
6Qwen 3.6 27B (Non-reasoning)23.73
7GPT-5.4 (Non-reasoning)22.74
8Claude Opus 4.622.33
9Claude Opus 4.522.13
10Qwen 3.6 35B A3B (Non-reasoning)20.4
11Llama 4 Scout13.59

Interactive version: theaggregate.ai/benchmark?slug=cdr-bench-order-agnostic-recipes · How It Works · Data refreshed daily, snapshot 2026-09-29.