List your company

Spatial Information on Spatial and Temporal Simulation Tasks Episode-level

3.543Subtask AccuracyCortex (full harness)
2.50822.776853.04553.31415Jul 6, 2026

Evaluation results

MethodLinks
Cortex (full harness)Level=Episode-level2026.073.5434.0447.587
Cortex (w/o harness)Level=Episode-level2026.073.0333.7716.805
Cortex (harness on skills)Level=Episode-level2026.073.0173.9586.985
GPT-5Level=Episode-level2026.072.924.47.321
GeminiLevel=Episode-level2026.072.8964.0336.929
Cortex (baseline)Level=Episode-level2026.072.5493.9696.518
Qwen3-VL-8B-InstructLevel=Episode-level2026.072.5483.926.468