| Ramain | Next best agent | |
|---|---|---|
| Online-Mind2Web | 99.3% | 99% (Aside) |
| WebVoyager | 99.2% | 99.2% (Polar) |
| WorkArena L3 | 89.4% | 11.5% (GPT-5) |
| WebArena-Verified Hard | 84.2% | 71.3% (Fable 5) |
Scores
We scored Ramain on four public benchmarks, strict pass or fail, against the best result published for any other agent.
Public benchmarks, % of tasks passedRamainNext best
OMW+0.3 pts
99.3
99 · Aside
VWTied
99.2
99.2 · Polar
WA L3+77.9 pts
89.4
11.5 · GPT-5
WAV Hard+12.9 pts
84.2
71.3 · Fable 5
OMW: Online-Mind2Web · VW: WebVoyager · WA L3: WorkArena L3 · WAV Hard: WebArena-Verified Hard
Where the gap opens
On everyday tasks on live websites, the leading agents are close. On multi-step work inside enterprise software, such as WorkArena L3 in ServiceNow, Ramain passes 89.4% of tasks where the next best result is 11.5%.