Nvidia open-sources Flora, a video‑labeling engine 125× cheaper than GPT‑6 Astra
Flora annotates robot video at $0.72 per hour, using three fine‑tuned Qwen3.5 models and claims higher quality than GPT‑6 Astra in tests.
Nvidia has released Flora, an open data engine for robot video labeling, claiming it can label an hour of video for 72 cents versus about $88 for OpenAI's GPT‑6 Astra.
In a test of 50 videos, Flora achieved 57.1 % annotation quality at $0.72 per hour, slightly ahead of GPT‑6 Astra’s 53.7 % quality at $88.49, while Claude Opus 5 scored lower and cost about $42 per hour.
Flora excels at detecting action boundaries and catching failed steps, reducing error rates by 27.8 % compared with Astra, though it sometimes over‑segments actions and mislabels them.
The engine is positioned as an open alternative to commercial services such as Scale AI’s Physical AI Data Engine and integrates with Nvidia’s broader robotics data pipeline.
Sources
- Nvidia open-sources Flora to make robot video labeling 125x cheaper than GPT-6 Astra, The Decoder, 25 Oct 2026
Written automatically from the sources above, then checked by our code: every number and name must appear in them, and no sentence may copy a source. Spotted a mistake? Report it.
Which AI can you use for free today?
Every free AI model, ranked by blind human votes, with the memory and speed it needs on your own GPU.