# Web Agent Navigation on WorkArena L2 147-task (test) Source: [Web Agent Navigation on WorkArena L2 147-task (test)]() Collection index: [llms.txt]() Data updated: 2026-04-10T08:05:12.628Z Evaluation of web agent performance on the specific L2 147-task test split of the WorkArena dataset, focusing on navigation success rate. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Web Agent Navigation | | Dataset (including any stated split) | WorkArena L2 147-task (test) | | Metric | Success Rate | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 10 of 10 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Success Rate](): Higher is better; unit: PERCENT. ## Recorded results | Method | Success Rate | Result date | Report paper | | --- | --- | --- | --- | | Gemini 3.1 Pro | 40 | Not recorded | 69d883f37d0172140ec77912 | | Claude 3.5 Sonnet | 38.8 | Not recorded | 69d883f37d0172140ec77912 | | Qwen3.5-27B | 18.9 | Not recorded | 69d883f37d0172140ec77912 | | o1-mini | 14.3 | Not recorded | 69d883f37d0172140ec77912 | | A3-Qwen3.5-9B | 9.7 | Not recorded | 69d883f37d0172140ec77912 | | Llama 3.1 405B | 8.9 | Not recorded | 69d883f37d0172140ec77912 | | GPT-4o | 7.5 | Not recorded | 69d883f37d0172140ec77912 | | Llama 3.1 70B | 3.4 | Not recorded | 69d883f37d0172140ec77912 | | Qwen3.5-9B (base) | 2.2 | Not recorded | 69d883f37d0172140ec77912 | | GPT-4o-mini | 2 | Not recorded | 69d883f37d0172140ec77912 | ### Gemini 3.1 Pro Success Rate: 40. Result ID: 69d8af1c47f34ce924c15e37. [Reporting paper]() Recorded evaluation configuration: {"Agent harness":"This work"}. This record does not assert independent reproduction. ### Claude 3.5 Sonnet Success Rate: 38.8. Result ID: 69d8af1c47f34ce924c15e31. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Agent harness":"BrowserGym ecosystem"}. This record does not assert independent reproduction. ### Qwen3.5-27B Success Rate: 18.9. Result ID: 69d8af1c47f34ce924c15e38. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"27B","Agent harness":"This work"}. This record does not assert independent reproduction. ### o1-mini Success Rate: 14.3. Result ID: 69d8af1c47f34ce924c15e32. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Agent harness":"BrowserGym ecosystem"}. This record does not assert independent reproduction. ### A3-Qwen3.5-9B Success Rate: 9.7. Result ID: 69d8af1c47f34ce924c15e39. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"9B","Agent harness":"This work"}. This record does not assert independent reproduction. ### Llama 3.1 405B Success Rate: 8.9. Result ID: 69d8af1c47f34ce924c15e34. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"405B","Agent harness":"BrowserGym ecosystem"}. This record does not assert independent reproduction. ### GPT-4o Success Rate: 7.5. Result ID: 69d8af1c47f34ce924c15e33. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Agent harness":"BrowserGym ecosystem"}. This record does not assert independent reproduction. ### Llama 3.1 70B Success Rate: 3.4. Result ID: 69d8af1c47f34ce924c15e35. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"70B","Agent harness":"BrowserGym ecosystem"}. This record does not assert independent reproduction. ### Qwen3.5-9B (base) Success Rate: 2.2. Result ID: 69d8af1c47f34ce924c15e3a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"9B","Agent harness":"This work"}. This record does not assert independent reproduction. ### GPT-4o-mini Success Rate: 2. Result ID: 69d8af1c47f34ce924c15e36. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Agent harness":"BrowserGym ecosystem"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.