OfficePrism-420 - ReAct (Search) - Strict Success: leaderboard

OfficePrism-420 (arXiv 2611.22262) single-domain enterprise agent evaluation (ReAct (Search) - Strict Success) across 420 tasks in 7 business functions.

Metric: Strict Success Rate (%) over 420 single-domain enterprise tasks. Source: arxiv.org. Saturation forecast: Around July 2027. 14 models tracked. Updated 22 Nov 2026.

Top models

#ModelScoreOverall rank
1Kimi K358.3#62
2Claude Sonnet 558.3#124
3GPT-5.558.1#56
4DeepSeek V4 Pro58.1#201
5GLM-5.258.1#154
6Claude Opus 4.757.9#86
7Claude Opus 4.857.9#64
8GLM-5.157.4#200
9DeepSeek V4 Flash56.2#273
10Qwen 3.7 Plus52.9#221
11Qwen 3.6 Max48.8#192
12Qwen 3.6 Plus48.6#245
13Kimi K2.647.9#197
14Qwen 3.7 Max47.6#160

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=officeprism-420-react-search-strict-success · How It Works · Data refreshed daily, snapshot 2026-11-22.