ScreenSpot Pro
ScreenSpot Pro - visual agent benchmark for GUI screen element grounding and interaction.
Category
General Agents
Score type
percent
Scale
relative
higher is better
Models tested
9
What is ScreenSpot Pro?
ScreenSpot Pro is a benchmark for how well a model acts as a general-purpose agent. The official description says: "ScreenSpot Pro - visual agent benchmark for GUI screen element grounding and interaction". Below we describe in simple words what benchmarks of this type generally examine.
What does it test?
ScreenSpot Pro measures how well a model acts as a general-purpose agent. In detail, it checks:
- solving long, multi-step tasks with many intermediate decisions
- using tools correctly (browsers, terminals, APIs, calendars, other apps)
- adapting when something unexpected happens mid-task
- staying on goal over long task horizons instead of wandering off
How to read the score
The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better.
Official methodology note: ScreenSpot Pro - visual agent benchmark for GUI screen element grounding and interaction.
Frequently asked questions
What is the ScreenSpot Pro benchmark?
ScreenSpot Pro is a benchmark for how well a model acts as a general-purpose agent. The official description says: 'ScreenSpot Pro - visual agent benchmark for GUI screen element grounding and interaction'. Below we describe in simple words what benchmarks of this type generally examine.
What does ScreenSpot Pro test?
ScreenSpot Pro mainly measures how well a model acts as a general-purpose agent. Concretely, it checks: solving long, multi-step tasks with many intermediate decisions; using tools correctly (browsers, terminals, APIs, calendars, other apps); adapting when something unexpected happens mid-task.
How is the ScreenSpot Pro score calculated?
The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better.
Where do I get ScreenSpot Pro data and scores?
The official benchmark page/reference is linked on this page. Every score on Open LLM Radar is also available as open data: download the scores JSON (http://open-llm-radar-api:8000/api/v1/benchmarks/screenspot-pro/scores) or pull it directly from the Open LLM Radar API with any HTTP client.
Benchmark data & downloads
Where to get, inspect and reuse the data behind this benchmark.
Benchmark dataset
The question/task data itself — download and evaluate locally
Benchmark metadata (JSON)
This page's benchmark record from the Open LLM Radar API
All model scores (JSON download)
Full score table as JSON; paginate with limit/offset params
Open LLM Radar API docs
Swagger UI to explore and reuse all radar API endpoints
Model scores 9 models tested
| # | Model | Score | Date | Source | Verified |
|---|---|---|---|---|---|
| 1 |
Qwen3.6-27B
Qwen Team (Alibaba Cloud)
|
76.1%
|
01.08.2026 | Meta Superintelligence Lab (official model card) | Unverified |
| 2 |
Gemma4-31B-it
Google
|
75.9%
|
01.08.2026 | Meta Superintelligence Lab (official model card) | Unverified |
| 3 |
Muse-Glimmer-30B
Meta Superintelligence Lab
|
75.4%
|
01.08.2026 | Meta Superintelligence Lab (official model card) | Verified |
| 4 |
Qwen3.5-122B-A10B
Qwen Team (Alibaba Cloud)
|
70.4%
|
01.02.2026 | Report | Verified |
| 5 |
Qwen3.5-27B
Qwen Team (Alibaba Cloud)
|
70.3%
|
01.02.2026 | Report | Unverified |
| 6 |
Qwen3.5-35B-A3B
Qwen Team (Alibaba Cloud)
|
68.6%
|
01.02.2026 | Report | Unverified |
| 7 |
Qwen3-235B-A22B
Qwen Team (Alibaba Cloud)
|
62%
|
01.02.2026 | Report | Unverified |
| 8 |
Qwen3.5-4B
Qwen
|
60.3%
|
01.02.2026 | Report | Verified |
| 9 |
LFM3-VL-2B
LiquidAI
|
55.5%
|
— | Report | Verified |