CITBench: leaderboard

Metric: Average performance (%; mean Level-Accuracy with partial credit, each of repeated runs scored 0 to 1, on the offline single-turn tabular data processing tasks of CITBench, a balanced 50% slice of its 1,296 instances in 18 matching, cleaning, augmentation and transformation task types; temperature 1.0, 8,192 output tokens; unweighted mean of the Base, Complex-Rule, Multi-Input and Multi-Step tiers). Source: arxiv.org. Saturation forecast: Around December 2026. 12 models tracked.

Top models

#ModelScore
1Claude Sonnet 4.570.33
2Gemini 3 Flash68.5
3Gemini 3 Pro65.32
4GLM-4.762.14
5GPT-5.261.74
6Claude Haiku 4.554.64
7DeepSeek V3.253.51
8GPT-4o52.35
9Qwen 3 Coder Flash49.56
10Qwen 3 235B A22B49.27
11Qwen 3 32B47.16
12Qwen 3 14B40.95

Interactive version: theaggregate.ai/benchmark?slug=citbench · How It Works · Data refreshed daily, snapshot 2026-09-26.