OfficePrism-420 - DeepAgents (Full) - Strict Success: leaderboard

OfficePrism-420 (arXiv 2611.22262) single-domain enterprise agent evaluation (DeepAgents (Full) - 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 August 2027. 14 models tracked. Updated 22 Nov 2026.

Top models

#ModelScoreOverall rank
1Kimi K358.8#62
2Claude Opus 4.858.1#64
3Claude Sonnet 558.1#124
4GPT-5.556.4#56
5Claude Opus 4.756.4#86
6GLM-5.256#154
7DeepSeek V4 Pro53.6#201
8DeepSeek V4 Flash53.6#273
9GLM-5.153.6#200
10Qwen 3.7 Max51.7#160
11Kimi K2.651.4#197
12Qwen 3.7 Plus51.4#221
13Qwen 3.6 Plus51.2#245
14Qwen 3.6 Max49.8#192

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-deepagents-full-strict-success · How It Works · Data refreshed daily, snapshot 2026-11-22.