Meta (Xuwei Ding et al.) · Spatial Action Benchmark
TuringSpace: Grounding Spatial Intelligence in Action with Interactive CAPTCHAs
TuringSpace is a synthetic web environment of 125 interactive CAPTCHAs covering five spatial abilities, solved through clicks and drags, with each task generating new instances, a checker and an oracle. Humans solve 94.1% of tasks, the best model GPT-6 Astra fails 25.9%, and the best open-weight model Kimi K3 fails 68.0%. Fine-tuning Qwen3.5-9B on the gym raises its success rate from 1.6% to 33.1%, with gains across seven out-of-distribution benchmarks, including 9.9 points on VSI-Bench indoor videos. The authors state they will release TuringSpace.
Open the source ↗Similar Research & Safety releases
- Self-Improving Memory
Meta FAIR · Self-Improving Memory · Week 48, 2026
- NovaAtom-Lite
Geodesic Intelligence · All-atom structure prediction · Week 47, 2026
- Sage-Taiyi: A Metacognitive Architecture for Scientific Discovery
Shanghai AI Laboratory (Weida Wang et al.) · Metacognitive Co-Scientist · Week 47, 2026
- LongBanana
Kolors Team, Kuaishou Technology (Haoxiang Cao et al.) · Image Synthesis Benchmark · Week 47, 2026
- Random Is Almost Enough? Stable and Low-Cost Benchmark Compression via BestRand
George Ling, Shanshan Zhong, Lijin Yang, Hao Yang, Nanxiang Li, Zhongzhan Huang (Carnegie Mellon University) · Benchmark Compression · Week 46, 2026
- Dream-RSI
Google, DeepMind, Maryland & Virginia · Recursive Self-Improvement · Week 39, 2026
Published by Quanta AI Labs, an AI agency in India. Browse every week in AI News Weekly.