ServiceNow
We track 5 benchmarks that ServiceNow made or helped make.
3 in use 4 with public code 5 with a leaderboard here
Benchmarks
| Name | Kind | Released | Best score | Status |
|---|---|---|---|---|
| WebArena-Verified Agents complete 812 audited WebArena tasks on self-hosted websites. Deterministic evaluators check the agent's answer and captured network traces. | Web browsing | Dec 2025 | | In use |
| EnterpriseOps-Gym Agents do 1,150 expert-written tasks in eight enterprise domains, such as IT, HR and customer service, using 512 tools over live MCP servers. SQL checks on the final database state score each task. | Apps and MCP servers | Mar 2026 | 45.9% | In use |
| EVA-Bench Voice agents hold simulated spoken conversations to complete enterprise tasks. Scores cover task success and voice-specific failure modes. | Speech recognition and translation | May 2026 | 0.74 | In use |
| CascadeBench Agents predict the cascade of changes that one action sets off in an enterprise system, where business rules differ between deployments. The environments are synthetic and follow the World of Workflows method. Tests include shifts in the system's rules to check whether a model reads the active configuration or relies on what it learned in training. | Interactive world models | May 2026 | 41.8% | Not in cards |
| SynthDocBench Vision-language models answer questions about synthetic long documents where length, layout, modality and difficulty are varied one at a time to locate failure causes. | Documents, charts and text in images | Jul 2026 | | Not in cards |