OccuBench - Healthcare & Life Sciences: leaderboard

Metric: Completion rate (%) on the Healthcare & Life Sciences 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
1Kimi K2.5 (Thinking)81
2Qwen 3.5 Plus (Thinking)81
3GPT-5.2 (High)76
4Claude Opus 4.6 (High)76
5Qwen 3.5 Flash (Thinking)76
6Claude Sonnet 4.5 (Thinking)71
7DeepSeek V3.2 (Thinking)71
8Claude Sonnet 4.6 (Thinking)67
9Gemini 3.1 Pro (Preview)62
10MiniMax-M2.757
11Claude Sonnet 4 (Thinking)57
12GLM-5 (Thinking)57
13Claude Opus 4 (Thinking)57
14Claude Opus 4.5 (Thinking)52

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