OccuBench - Commerce & Consumer: leaderboard

Metric: Completion rate (%) on the Commerce & Consumer industry's tasks in the clean environment (E0); 382 tool-use task instances from 100 professional scenarios, the environment simulated by Gemini-3-Flash-Preview and each trajectory passed or failed by a rubric verifier; thinking mode on, high reasoning effort where configurable; higher is better. Source: arxiv.org. Saturation forecast: Not forecast. 15 models tracked.

Top models

#ModelScore
1Qwen 3.5 Plus (Thinking)81
2Gemini 3.1 Pro (Preview)75
3Claude Sonnet 4.5 (Thinking)69
4DeepSeek V3.2 (Thinking)67
5GPT-5.2 (High)67
6GLM-5 (Thinking)67
7Qwen 3.5 Flash (Thinking)67
8Claude Sonnet 4.6 (Thinking)64
9Claude Sonnet 4 (Thinking)61
10MiniMax-M2.756
11Kimi K2.5 (Thinking)56
12Claude Opus 4.5 (Thinking)56
13Claude Opus 4.6 (High)53
14Claude Opus 4 (Thinking)50

Interactive version: theaggregate.ai/benchmark?slug=occubench-commerce-consumer · How It Works · Data refreshed daily, snapshot 2026-10-07.