WebArena C
AI benchmark · Agentic
What is WebArena?
WebArena measures whether autonomous language agents can complete diverse, long-horizon tasks on self-hosted, fully functional websites, scored by functional correctness of task completion.
Computer use mostly
“we focus on agents that perform tasks on the web, and create an environment with fully functional websites from four common domains”
arxiv.org/abs/2307.13854 ↗“e-commerce, social forum discussions, collaborative software development, and content management”
arxiv.org/abs/2307.13854 ↗
Paper: arxiv.org/abs/2307.13854 ↗
Lineage
Benchmarks this one is built on and built from: versions, subsets, extensions and the indexes that combine them.
Derived benchmarks
Executable Benchmarking Suite for Tool-Using Agents extension
“An executable benchmarking suite that evaluates closed-loop tool-using agents across web (WebArena Verified, MiniWoB++) and code (SWE-Gym/SWE-bench) workloads under a shared evidence-admission contract, recording latency, invalid actions, and patch-generation cost.”
WebArena-Verified extension
WebChoreArena extension
“WebChoreArena measures how well web browsing agents complete 532 tedious, labor-intensive web chores requiring massive memory, calculation, and long-term memory across reproducible WebArena sites.”
Tier C
Niche: a handful of models scored, or built by an RL vendor.
How much stands behind a tier: usage, who backs it, a confirmed line from its own site or paper, and a clear read on X. High with three of the four, medium with two, low otherwise.
- 4 models scored
- fewer than 5 models scored
- in curated RL directories (2 pts)
On X · well regarded, moderate buzz, from 9 posts about WebArena in the last month
WebArena and its Verified variant are widely cited in model launches from Liquid AI, NVIDIA, and Salesforce, with some research noting run-to-run variance and task-order fragility.
How we got to tier C
- Adoption: 4 models scored; fewer than 5 models scored. Tier D.
- On X: well regarded, moderate buzz. “WebArena and its Verified variant are widely cited in model launches from Liquid AI, NVIDIA, and Salesforce, with some research noting run-to-run variance and task-order fragility.” No move.
- In curated RL directories: 2 points, a floor of C. Raised to C.
- Confirmed: a line from its own site or paper places it.
A clear view on X moves a tier one step up or down; most entries don't move. How tiers work.
Where WebArena sits
Its domain on the map, live. Hover a logo for its name; click to open it. Open it on the full map
Next: browse all benchmarks. Entry from the RL Research daily scrape of public sources, 2026-11-05 snapshot.