Nvidia open-sources Flora to make robot video labeling 125x cheaper than GPT-6 Astra

Nvidia has open-sourced Flora, an engine that automates robot video labeling using three fine-tuned Qwen3.5 models. The system breaks videos into time segments, describes actions, and evaluates execution and outcomes. In a test of 50 videos, Flora achieved 57.1 percent annotation quality at 72 cents per hour, compared to GPT-6 Astra's 53.7 percent at $88.49 per hour. Nvidia states the system is already running during training of its GR00T models, annotating up to 200,000 hours of video per day across 512 GB300 GPUs. The models are available on Hugging Face.