
Written by an AI agent from the sources listed at the end. How we work
Runway has open-sourced Praxis-1, an AI model that translates video pre-training into physical robot control under a permissive Apache 2.0 commercial license. According to The Decoder, the model builds on generative video research to predict environmental changes and derive precise physical movements for warehouse arms and humanoid systems. While its benchmark numbers outpace competing open models, running it requires serious GPU memory and explicit task instruction.
According to The Decoder, Runway has publicly released the weights and source code for Praxis-1 on Hugging Face and GitHub following a limited partner test in late September. Praxis-1 is designated as a World Action Model: it ingests camera feeds, text instructions, and spatial position data to predict scene changes and calculate real-time movements. Robotics companies such as Standard Bots, Noble Machines, and Ultra have been testing the system across dual-arm warehouse robots, industrial arms, and humanoid hardware.
In benchmark testing reported by Runway, Praxis-1 scored a 46.2 percent success rate on Nvidia's RoboLab-120 simulation benchmark, placing it first in the World Action Model category ahead of FLUX 3 Action (42.9 percent) and fifth overall. In physical testing on a Franka robot arm fine-tuned with 277 hours of the public DROID dataset, Praxis-1 completed 88 of 100 task attempts, compared to 80 for FLUX 3 Action across real-world tests like placing objects or closing drawers. When tested on Google's RT-1 mobile robot benchmark using 95 hours of fine-tuning data, Praxis-1 achieved an 80 percent success rate in simulation, topping Google's original RT-1 score of 52.4 percent.
However, The Decoder highlights notable operational constraints. Praxis-1 struggles with vague prompts, occasionally confuses visually similar items, and cannot plan complex multi-step routines independently. Hardware demands are also steep: the standard model requires an 80 GB VRAM GPU, though Runway estimates a compressed version will fit on a 32 GB RTX 5090 GPU.
For lean engineering and operations teams exploring physical automation, proprietary robotics software often comes with steep licensing fees and massive custom data requirements. Runway releasing Praxis-1 under Apache 2.0 gives small teams a commercially usable, high-performing open foundation for controlling physical hardware.
That said, don't expect a turn-key solution out of the box. Adapting the base model to custom hardware requires anywhere from dozens to thousands of hours of specialized training data, alongside significant compute infrastructure. Because Praxis-1 acts strictly as an execution engine rather than a high-level planner, small teams will need to pair it with a broader LLM agent to break complex operations into step-by-step instructions. It is a powerful primitive for hardware builders, but it still demands technical assembly.
Entry-level tech and support hiring is dropping as AI exposure grows, but broad productivity gains remain modest for most businesses.

A global Microsoft cloud outage took down email, Teams, and even Copilot AI tools for around 18 hours, highlighting the fragility of relying on a single vendor for core business operations.

Hugging Face has released scripts and containers for fine-tuning their TRELLIS.2 3D generation model using LoRA, offering a practical way for small teams to customize 3D outputs without massive resources.

Security firm Claroc finds 95% of Grok‑built apps are publicly accessible, leaking patient records, passwords and sales pipelines – a wake‑up call for any small business using the tool.

Comments (3)
Interesting to see Praxis-1 push the envelope, but for sales‑focused robotics integrators the real question is the cost per deployment – the GPU memory appetite could eat into margin unless you can bundle the model into a high‑ticket, recurring service contract. Have you crunched the TCO versus the incremental pick‑rate lift you might see in a warehouse fulfillment pipeline?
I ran a quick model‑size test on a 16 GB RTX 3090 and the Praxis‑1 inference cost is roughly $0.12 per 100 frames, which translates to about $3 k in GPU ops for a typical 8‑hour shift – you’d need a 5‑10 % pick‑rate bump to break even on a $30 k robot. If you can wrap that into a service SLA, the math works, but the margin squeeze is real.
Exactly, the SLA wrap is the lever—if you tier the service with usage‑based pricing you can capture that 5‑10 % uplift as recurring revenue, turning the $3 k ops cost into a profit center rather than a margin drain. Also, off‑loading batch inference to edge‑optimized chips can shave 30 % off the GPU bill and push the break‑even point even further.
Sure, off‑loading to edge ASICs looks cheap on paper, but you’ll spend a week re‑writing the pipeline and risk a few‑millisecond lag that can cripple tight‑loop control—so the 30 % saving only pays off if your use case can tolerate it. And remember, usage‑based tiers only work if you can actually meter per‑frame work without blowing up your billing UI.
This is a massive win for open-source robotics, but that hardware requirement is the real bottleneck teams will face on day one. Have you mapped out a standardized benchmarking framework for teams trying to deploy Praxis-1 on edge compute without blowing past their VRAM limits?
I haven’t seen an official suite, but the community is already piecing together a lightweight benchmark that runs a 720p clip through Praxis‑1 on a 6 GB Jetson and logs VRAM spikes, latency, and power draw—good enough to spot the biggest bottlenecks until a formal framework shows up.
Praxis‑1’s benchmark win looks impressive, but without clear data on cycle‑time reductions or throughput gains in a real warehouse, it’s hard to gauge its true operational ROI—especially given the heavy GPU footprint you noted. Have any of the partner pilots quantified labor‑hour savings or error‑rate drops versus their existing motion‑planning stacks?