SpaceDG: leaderboard
Metric: Avg. (self-reported). Source: benchmarklist.com. 30 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 56.7 |
| 2 | Gemini 3.1 Flash Lite | 48.8 |
| 3 | GPT-5.4 | 46.2 |
| 4 | Qwen 3.6 Plus | 46 |
| 5 | Qwen 3 VL 32B Instruct | 45.2 |
| 6 | Claude Sonnet 4.6 | 44.2 |
| 7 | Qwen 3.5 27B | 43.8 |
| 8 | Qwen 3.6 35B A3B | 42.4 |
| 9 | Qwen 3 VL 8B Instruct | 42.1 |
| 10 | Qwen 3.5 9B | 41.7 |
| 11 | Qwen 3.5 4B | 40.4 |
| 12 | InternVL3-8B | 40.1 |
| 13 | Qwen 3 VL 4B Instruct | 39.7 |
| 14 | Grok 4.1 Fast | 35.6 |
| 15 | Llama 4 Maverick | 34.3 |
Interactive version: theaggregate.ai/benchmark?slug=spacedg · How It Works · Data refreshed daily, snapshot 2026-09-05.