OccuBench - Public Service & Governance: leaderboard

Metric: Completion rate (%) on the Public Service & Governance 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: Around January 2027. 15 models tracked.

Top models

#ModelScore
1GPT-5.2 (High)84
2Claude Sonnet 4 (Thinking)76
3Qwen 3.5 Plus (Thinking)76
4Claude Opus 4 (Thinking)76
5Gemini 3.1 Pro (Preview)72
6DeepSeek V3.2 (Thinking)72
7Kimi K2.5 (Thinking)72
8Claude Opus 4.5 (Thinking)72
9Claude Opus 4.6 (High)68
10GLM-5 (Thinking)68
11Qwen 3.5 Flash (Thinking)68
12Claude Sonnet 4.6 (Thinking)64
13MiniMax-M2.760
14Claude Sonnet 4.5 (Thinking)60

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