Benchmark
ScreenSpot Pro Benchmark
What is ScreenSpot Pro?
ScreenSpot-Pro is a novel GUI grounding benchmark designed to rigorously evaluate the grounding capabilities of multimodal large language models (MLLMs) in…
- Released
- 2025-04-04
- Modality
- multimodal
- Evaluates
- multimodal, spatial reasoning, grounding, vision
- Openness
- unknown
- Importer
- LLM Stats
- Reported scores
- 25
Source provenance
- Original evidence https://api.zeroeval.com/leaderboard/benchmarks/screenspot-pro?top_n=500
ScreenSpot Pro paper and code
ScreenSpot Pro results and reported scores
LLM Stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Claude Opus 4.8 | Anthropic | 0.879 | 2026-05-28 | evidence |
| Gemini 3 Flash | 0.691 | 2025-12-17 | evidence | |
| Gemini 3 Pro | 0.727 | 2025-11-18 | evidence | |
| GPT-5.2 | OpenAI | 0.863 | 2025-12-11 | evidence |
| Muse Glimmer-30B | Meta | 0.754 | 2026-08-10 | evidence |
| Muse Spark | Meta | 0.841 | 2026-04-08 | evidence |
| Qwen2.5 VL 32B Instruct | Qwen | 0.394 | 2025-02-28 | evidence |
| Qwen2.5 VL 72B Instruct | Qwen | 0.436 | 2025-01-26 | evidence |
| Qwen2.5 VL 7B Instruct | Qwen | 0.29 | 2025-01-26 | evidence |
| Qwen3 VL 235B A22B Instruct | Qwen | 0.62 | 2025-09-22 | evidence |
| Qwen3 VL 235B A22B Thinking | Qwen | 0.618 | 2025-09-22 | evidence |
| Qwen3 VL 30B A3B Instruct | Qwen | 0.605 | 2025-09-22 | evidence |
| Qwen3 VL 30B A3B Thinking | Qwen | 0.573 | 2025-09-22 | evidence |
| Qwen3 VL 32B Instruct | Qwen | 0.579 | 2025-09-22 | evidence |
| Qwen3 VL 32B Thinking | Qwen | 0.571 | 2025-09-22 | evidence |
| Qwen3 VL 4B Instruct | Qwen | 0.595 | 2025-09-22 | evidence |
| Qwen3 VL 4B Thinking | Qwen | 0.492 | 2025-09-22 | evidence |
| Qwen3 VL 8B Instruct | Qwen | 0.546 | 2025-09-22 | evidence |
| Qwen3 VL 8B Thinking | Qwen | 0.466 | 2025-09-22 | evidence |
| Qwen3.5-122B-A10B | Qwen | 0.704 | 2026-02-24 | evidence |
| Qwen3.5-27B | Qwen | 0.703 | 2026-02-24 | evidence |
| Qwen3.5-35B-A3B | Qwen | 0.686 | 2026-02-24 | evidence |
| Qwen3.6 Plus | Qwen | 0.682 | 2026-04-02 | evidence |
| Qwen3.7-Plus | Qwen | 0.79 | 2026-05-31 | evidence |
| Qwen3.8 Max | Qwen | 0.845 | 2026-08-02 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.