APEX-Agents-AA

An implementation of the APEX-Agents benchmark, testing AI agents on long-horizon, cross-application tasks in professional-services environments with realistic application tooling.

Models: 9 current, 33 in total
Median: 27.4%
Range: 0.7%–42.4%

Score distribution

Lowest: 0.7%
Median: 27.4%
Highest: 42.4%

Leaderboard

Include deprecated models

Rank	Model	Score
1	Qwen3.8 Max (0902) Alibaba	42.4%
2	Step 5 Preview StepFun	38.0%
3	Gemini 3.1 Pro Preview Google	32.0%
4	Apodex 1.1 Apodex	31.2%
5	Ling-3.0-flash-Fin InclusionAI	27.4%
6	Qwen3.7 Plus Alibaba	22.4%
7	Qwen3.5 397B A17B (Reasoning) Alibaba	15.3%
8	gpt-oss-120b (High) OpenAI	3.1%
9	gpt-oss-20b (High) OpenAI	0.7%

1
Qwen3.8 Max (0902)
Alibaba
42.4%
2
Step 5 Preview

Data from Artificial Analysis

Command Palette

APEX-Agents-AA

Score distribution

Leaderboard