AgentWorldBench Overall
AgentWorldBench Overall - mean across all 7 agent interaction domains (MCP, Search, Terminal, SWE, Android, Web, OS). Five-dimensional rubric, normalized to 0-100.
Category
General Agents
Max Score
100.0
Score Type
percent
Active
Yes
Model Scores
Scores for AgentWorldBench Overall
| Model | Score | Date | Verified |
|---|---|---|---|
| Qwen-AgentWorld-35B-A3B |
81.57%
|
24.06.2026 | Verified |
| Qwen-AgentWorld-397B-A17B |
100.00%
|
24.06.2026 | Unverified |
| Qwen3.5-35B-A3B |
12.79%
|
24.06.2026 | Unverified |
| Qwen3.5-397B-A17B |
68.47%
|
24.06.2026 | Unverified |
| Qwen3.6-Plus |
37.25%
|
24.06.2026 | Unverified |
| GPT-5.4 |
96.35%
|
24.06.2026 | Unverified |
| Claude Opus 4.8 |
83.16%
|
24.06.2026 | Unverified |
| Claude Opus 4.6 |
92.77%
|
24.06.2026 | Unverified |
| Gemini 3.1 Pro |
67.12%
|
24.06.2026 | Unverified |
| Claude Sonnet 4.6 |
78.79%
|
24.06.2026 | Unverified |
| DeepSeek-V4-Pro |
54.41%
|
24.06.2026 | Unverified |
| GLM-5.1 |
41.22%
|
24.06.2026 | Unverified |
| Kimi K2.6 |
57.98%
|
24.06.2026 | Unverified |
| MiniMax-M2.7 |
46.12
|
24.06.2026 | Unverified |