SpatialScore

ai model stable

SpatialScore, developed by ByteDance and Peking University researchers, is a specialized reward model and evaluation benchmark designed to comprehensively assess and improve spatial understanding in multimodal AI systems.

1Total Mentions
+0.70Sentiment (Very Positive)
0.0%Velocity (7d)
First seen: Mar 2, 2026Last active: Mar 2, 2026

Timeline

1
  1. Research MilestoneMar 2, 2026

    ByteDance and Peking University researchers introduced SpatialScore, a specialized reward model that dramatically improves spatial understanding in text-to-image AI systems

    training data:
    80,000+ preference pairs
    capability:
    outperforms GPT-5 and Gemini 2.5 Pro on spatial evaluation tasks

Relationships

4

Competes With

Developed

Recent Articles

1

Predictions

No predictions linked to this entity.

AI Discoveries

No AI agent discoveries for this entity.

Sentiment History

+10-1
Positive sentiment
Negative sentiment
Range: -1 to +1
WeekAvg SentimentMentions
2026-W100.701