WebArena-Verified
Can the model plan, use tools, and finish a multi-step task?
Agents5 ranked models5 reported values1 reportsHigher is better
Top rankings
One row per model, using its best reported score across effort settings.
| Rank | Model | Best score | Best reported setting |
|---|---|---|---|
| 1 | Claude Opus 4.8Anthropic | 71.2% | max effortMeta report1 value · 1 reportJul 9, 2026 · source |
| 2 | Gemini 3.1 ProGoogle | 69.0% | high effortMeta report1 value · 1 reportJul 9, 2026 · source |
| 3 | Muse Spark 1.1Meta | 69.0% | xhigh effortMeta report1 value · 1 reportJul 9, 2026 · source |
| 4 | GPT-5.5OpenAI | 67.0% | xhigh effortMeta report1 value · 1 reportJul 9, 2026 · source |
| 5 | Muse Spark 1.0Meta | 59.0% | Reported configurationMeta report1 value · 1 reportJul 9, 2026 · source |
Effort curve
Every sourced cost-linked effort value for this exact version. Lines connect complete sweeps only.
No cost-linked effort sweep for this version.
Definition and comparison boundarypublic methodology
Can the model plan, use tools, and finish a multi-step task?
Agentic execution across tools, environments, or long-running workflows. Higher is better. The value is the percentage reported in this lab's table.
Harness, tools, prompts, attempt count, and benchmark version can materially change the result. Compare within one reporting context.
812 tasks, OpenClaw browser tool, 150-call cap, mean pass@1.
Method / source