{"version":1,"type":"story","url":"https://digestai.news/story/nvidia-open-sources-flora-for-robot-video-labeling","json":"https://digestai.news/story/nvidia-open-sources-flora-for-robot-video-labeling.json","markdown":"https://digestai.news/story/nvidia-open-sources-flora-for-robot-video-labeling.md","slug":"nvidia-open-sources-flora-for-robot-video-labeling","headline":"Nvidia open-sources Flora for robot video labeling","summary":"Nvidia has released Flora, an open-source data engine designed to label robot training videos. The system uses three fine-tuned Qwen3.5 models to generate action descriptions, progress ratings, and success judgments. Nvidia estimates that labeling one hour of video costs 72 cents with Flora, compared to roughly $88.50 using GPT-6 Astra. These cost figures are projections based on an assumed GPU rental price of $10 per hour, not production measurements.\n\nIn tests, Flora outperformed GPT-6 Astra in detecting action boundaries and identifying mid-episode errors. On RoboRewardBench, Flora’s error rate was 27.8 percent lower than Astra’s. However, Astra often provided more precise descriptions and led in rubric-based scoring. Nvidia notes that Flora is already running in the training of its GR00T models, processing up to 200,000 hours of video per day across 512 GB300 GPUs.\n\nThe release positions Flora as an open alternative to commercial services like Scale AI’s Physical AI Data Engine. It aims to help researchers scale robotics datasets and test approaches like transferring human demonstrations to robots. This move supports Nvidia’s broader strategy to cover the entire robotics value chain, from hardware to simulation and AI models, while driving demand for its chips.","keyPoints":["Nvidia open-sources Flora, using three fine-tuned Qwen3.5 models for robot video labeling.","Nvidia estimates Flora costs 72 cents per hour of video, versus $88.50 for GPT-6 Astra.","Flora detects mid-episode errors better than competitors, with a 27.8% lower error rate than Astra."],"whyItMatters":"Flora reduces the cost of labeling robot training data, making it easier for developers to scale robotics datasets. It strengthens Nvidia’s ecosystem by providing a tool that drives demand for its GPUs while competing with commercial data services.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":["Nvidia","OpenAI","Scale AI","Google"],"models":["Flora","GPT-6 Astra","Qwen3.5","GR00T","Claude Opus 5","Gemini 3.8 Flash"],"people":[]},"firstPublishedAt":"2026-10-25T10:08:49Z","updatedAt":"2026-10-25T10:08:49Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"The Decoder","title":"Nvidia open-sources Flora to make robot video labeling 125x cheaper than GPT-6 Astra","url":"https://the-decoder.com/nvidia-open-sources-flora-to-make-robot-video-labeling-125x-cheaper-than-gpt-6-astra","publishedAt":"2026-10-25T10:08:49Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Nvidia open-sources Flora for robot video labeling\", 25 October 2026, https://digestai.news/story/nvidia-open-sources-flora-for-robot-video-labeling","publisher":"Digest AI","title":"Nvidia open-sources Flora for robot video labeling","datePublished":"2026-10-25T10:08:49Z","url":"https://digestai.news/story/nvidia-open-sources-flora-for-robot-video-labeling"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}