Overview of NVIDIA 'Flora' - A Suite of Open Models for Automatically Generating Robot Learning Data
Overview of NVIDIA 'Flora' - A Suite of Open Models for Automatically Generating Robot Learning Data
NVIDIA has released 'Flora', a suite of open AI models designed to automatically generate robot learning data. Flora includes three models for action segmentation, captioning, and reward generation, and is claimed to outperform other models in certain benchmarks while being significantly more cost-effective for large-scale video processing.
NVIDIA's Flora is presented as an open-source tool for generating robot learning data, with performance benchmarks against models like GPT-6 Astra and Claude Opus 5. It includes three models for action segmentation, captioning, and reward generation, and is claimed to be cost-effective for large-scale video processing.
Source: Robotics — Robot learning news discovery · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
video hours processed per day
200,000 hours
View original evidence
at approximately 1/125th the cost of GPT-6 Astra. Furthermore, it has a track record of labeling 200,000 hours of video per day using 512 GB300 GPUs. This demonstrates that robot learning data can be generated not just at the laboratory level, but at the scale of large data centers. However, these costs and processing speeds are based on NVIDIA's measurement conditions. 6. Model release The pre-trained weights for Fl
Open source S19cost per hour of video
0.72
View original evidence
criteria. These are the benchmark results reported by NVIDIA. 5. Processing cost and scalability A major feature of Flora is that it achieves evaluation performance comparable to frontier models at a low cost. The labeling cost per hour of video estimated in the paper is as follows: ・Flora: $0.72 ・GPT-6 Astra: $88.49 ・Claude Opus 5: $41.70 ・Gemini 3.8 Flash: $8.48 Under NVIDIA's evaluation conditions, Flora operates
Open source S18cost per hour of video
88.49
View original evidence
criteria. These are the benchmark results reported by NVIDIA. 5. Processing cost and scalability A major feature of Flora is that it achieves evaluation performance comparable to frontier models at a low cost. The labeling cost per hour of video estimated in the paper is as follows: ・Flora: $0.72 ・GPT-6 Astra: $88.49 ・Claude Opus 5: $41.70 ・Gemini 3.8 Flash: $8.48 Under NVIDIA's evaluation conditions, Flora operates
Open source S18cost per hour of video
41.7
View original evidence
criteria. These are the benchmark results reported by NVIDIA. 5. Processing cost and scalability A major feature of Flora is that it achieves evaluation performance comparable to frontier models at a low cost. The labeling cost per hour of video estimated in the paper is as follows: ・Flora: $0.72 ・GPT-6 Astra: $88.49 ・Claude Opus 5: $41.70 ・Gemini 3.8 Flash: $8.48 Under NVIDIA's evaluation conditions, Flora operates
Open source S18cost per hour of video
8.48
View original evidence
criteria. These are the benchmark results reported by NVIDIA. 5. Processing cost and scalability A major feature of Flora is that it achieves evaluation performance comparable to frontier models at a low cost. The labeling cost per hour of video estimated in the paper is as follows: ・Flora: $0.72 ・GPT-6 Astra: $88.49 ・Claude Opus 5: $41.70 ・Gemini 3.8 Flash: $8.48 Under NVIDIA's evaluation conditions, Flora operates
Open source S18
- code: not_reported
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
at approximately 1/125th the cost of GPT-6 Astra. Furthermore, it has a track record of labeling 200,000 hours of video per day using 512 GB300 GPUs. This demonstrates that robot learning data can be generated not just at the laboratory level, but at the scale of large data centers. However, these costs and processing speeds are based on NVIDIA's measurement conditions. 6. Model release The pre-trained weights for Fl
Open source S19
criteria. These are the benchmark results reported by NVIDIA. 5. Processing cost and scalability A major feature of Flora is that it achieves evaluation performance comparable to frontier models at a low cost. The labeling cost per hour of video estimated in the paper is as follows: ・Flora: $0.72 ・GPT-6 Astra: $88.49 ・Claude Opus 5: $41.70 ・Gemini 3.8 Flash: $8.48 Under NVIDIA's evaluation conditions, Flora operates
Open source S18
Source：Robotics — Robot learning news discovery · note.com