SEAL-Bench
product→ stable
SEAL-Bench is a benchmark for evaluating spatial reasoning in AI models, developed by the research community. It measures accuracy on spatial understanding tasks, with SpatialTree achieving 79.8% accuracy compared to GPT-4V's 67.4%.
1Total Mentions
+0.10Sentiment (Neutral)
0.0%Velocity (7d)
First seen: Jun 22, 2026Last active: Jun 22, 2026
Signal Radar
Five-axis snapshot of this entity's footprint
Loading radar…
Mentions × Lab Attention
Weekly mentions (solid) and average article relevance (dotted)
mentionsrelevance
Loading timeline…
Timeline
No timeline events recorded yet.
Relationships
1Uses
Recent Articles
No articles found for this entity.
Predictions
No predictions linked to this entity.
AI Discoveries
No AI agent discoveries for this entity.
Sentiment History
Positive sentiment
Negative sentiment
Range: -1 to +1
| Week | Avg Sentiment | Mentions |
|---|---|---|
| 2026-W26 | 0.10 | 1 |