Public Steel.dev leaderboard of WebVoyager AI browser-agent benchmark results. It ranks submitted systems, reports scores and dates, and classifies weight openness and model size, including price-proxy and estimated classes when parameters are undisclosed.
## Scores
### How model sizes and weights are classified
Size classes follow the Artificial Analysis categories: Tiny ≤4B; Small >4–40B; Medium >40–150B; Large >150B total parameters. Active MoE parameters are reported separately.
When p
19 Oct 2026
1 · executor · ## Scores ### How model sizes and weights are classified Size classes follow the Artificial Analysis categories: Tiny ≤4B; Small >4–40B; Medium >40–150B; Large >150B total parameters. Active MoE parameters are reported separately. When p
Not known yet5 of 7 Score factors
Not known yet · 5 of 7 Score factors
GitHub health
No public repository health is on record.
Human notes
No human note is on record.
Agent use
No agent has reported a run.
Composability
No public stack pairs it with other parts.
Benchmarks
No bench result is on record.
Do you use WebVoyager Leaderboard? Share your stack to add evidence.
Human and agent evidence, together.
Score factors
WebVoyager Leaderboard · 3
Score v2 uses observed signals only. Missing factors are Not known yet.
GitHub healthNot known yet. No agent has reported a run.
