LLM Stats (ScreenSpot Pro): leaderboard
GUI element grounding for professional desktop applications (IDEs, CAD tools, spreadsheets), testing ability to locate UI elements from natural language.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around January 2027. 26 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra | 92.7 | #6 |
| 2 | Qwen 3.8 Max | 84.5 | #35 |
| 3 | Qwen 3.7 Plus | 79 | #92 |
| 4 | Muse Glimmer 30B | 75.4 | #178 |
| 5 | Gemini 3 Pro | 72.7 | #80 |
| 6 | Qwen 3.5 122B A10B | 70.4 | #173 |
| 7 | Qwen 3.5 27B | 70.3 | #187 |
| 8 | Gemini 3 Flash | 69.1 | #99 |
| 9 | Qwen 3.5 35B A3B | 68.6 | #256 |
| 10 | Qwen 3.6 Plus | 68.2 | #120 |
| 11 | Qwen 3 VL 235B A22B Instruct | 62 | #273 |
| 12 | Qwen 3 VL 235B A22B (Thinking) | 61.8 | #238 (Qwen 3 VL 235B A22B) |
| 13 | Qwen 3 VL 30B A3B Instruct | 60.5 | #384 |
| 14 | Qwen 3 VL 4B Instruct | 59.5 | #515 |
| 15 | Qwen 3 VL 32B Instruct | 57.9 | #283 |
No result here: #3 Claude Opus 5.5, #7 GPT-6.1 Sol, #8 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-screenspot-pro · How It Works · Data refreshed daily, snapshot 2026-10-20.