Benchmark Almanac

WorkArena

GPT-4o (snapshot unspecified) 42.7% tops its leaderboard
Not in model cards

Web agents do routine knowledge-work tasks in a live ServiceNow instance, such as filling forms and filtering lists. WorkArena++ adds 682 compositional tasks at levels 2 and 3.

Level 1 has 33 atomic tasks with 19,912 instances. It runs through BrowserGym and AgentLab. WorkArena++ (NeurIPS 2024) scores are reported per level.

Released
March 2024
Kind
Web browsing
Level
University
Scored
Checked automatically
Answers
Agent task
Tasks
All public

In the authors’ words

We study the use of large language model-based agents for interacting with software via web browsers. Unlike prior work, we focus on measuring the agents' ability to perform tasks that span the typical daily work of knowledge workers utilizing enterprise software systems. To this end, we propose WorkArena, a remote-hosted benchmark of 33 tasks based on the widely-used ServiceNow platform. We also introduce BrowserGym, an environment for the design and evaluation of such agents, offering a rich set of actions as well as multimodal observations. Our empirical evaluation reveals that while current agents show promise on WorkArena, there remains a considerable gap towards achieving full task automation. Notably, our analysis uncovers a significant performance disparity between open and closed-source LLMs, highlighting a critical area for future exploration and development in the field.

Abstract of WorkArena: How Capable Are Web Agents at Solving Common Knowledge Work Tasks? (arXiv:2403.07718, CC0)
3
Entries on its own leaderboard, 0 models we track

GPT-4o (snapshot unspecified) tops the leaderboard with 42.7%

Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders.

  1. 42.7 GPT-4o (snapshot unspecified)
  2. 17.9 Llama 3 (size unspecified)
  3. 6.10 GPT-3.5 (snapshot unspecified)
  1. 1 GPT-4o (snapshot unspecified) accessibility tree 42.7%
  2. 2 Llama 3 (size unspecified) accessibility tree 17.9%
  3. 3 GPT-3.5 (snapshot unspecified) accessibility tree 6.10%

Source: Official WorkArena leaderboard

#EntryScoreSettingDateSource
1 GPT-4o (snapshot unspecified) 42.7% accessibility tree, BrowserGym baseline agent Official leaderboard
2 Llama 3 (size unspecified) 17.9% accessibility tree, BrowserGym baseline agent Official leaderboard
3 GPT-3.5 (snapshot unspecified) 6.10% accessibility tree, BrowserGym baseline agent Official leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Computer and browser use

All 69 in Computer and browser use
  • BrowseComp In use Hard-to-find facts that need persistent web browsing. Short verifiable answers graded for exact correctness. Claude Fable 5 93.3%
  • ScreenSpot-Pro In use GUI grounding on high-resolution professional-software screenshots. The model must locate the target UI element. GPT-6 Astra 92.7%
  • AndroidWorld In use Agents complete 116 tasks across 20 real Android apps in an emulator. Each task checks the device's system state to judge success. GPT-6 Astra 100%
  • OSWorld 2.1 In use Agents complete 108 long-horizon workflows across 31 self-hosted websites and desktop apps, averaging 318 tool calls. Scored by binary completion and checkpoint partial credit. Claude Opus 5.5 81.8%
  • WebArena-Verified In use Agents complete 812 audited WebArena tasks on self-hosted websites. Deterministic evaluators check the agent's answer and captured network traces. Muse Spark 1.1 69%
  • MobileWorld In use Agents complete 201 tasks across 20 mobile apps, including tasks that need user clarification or MCP tools. Backend checks verify success. GPT-5 55.3%