# Flux 3 Action Tops RoboLab, but Large Companies Can't Use It in a Product > Source: > Published: 2026-09-24 20:07:13+00:00 ## **FLUX 3 Action scores 42.9% on RoboLab, and its faster version scores 38.3%** Black Forest Labs released [FLUX 3 Action](https://bfl.ai/models/flux-3-action) on 22 September. It is a 7B model that predicts robot actions and future video frames, and its [weights are open](https://huggingface.co/black-forest-labs/flux-3-action-droid). NVIDIA’s [RoboLab leaderboard](https://research.nvidia.com/labs/srl/projects/robolab/leaderboard.html) ranks it first at 42.9%. That score comes from the slower of its two main versions. The faster version, the one BFL says runs 1.34 to 2.28 times faster than π0.5, scores 38.3%. HiDream-O1-Embodied, Atomic-WAM, and OASIS all score higher. The [license](https://huggingface.co/black-forest-labs/flux-3-action-base/blob/main/LICENSE.md) allows commercial use only by companies with under $5 million in annual revenue, affiliates included. Everyone else can use it only for non-commercial research. It also bars using the model’s outputs to train a similar model, so you cannot distill from it either. ## **Correction** Last week I wrote that SolarWM’s staged checkpoints above the 33B base were not out. That was wrong. [All three stages](https://huggingface.co/junchaoh-cs/SolarWM-H3-33B) were published on 13 September. On 22 September, the authors replaced the Stage2 checkpoint, so if you downloaded it before then, download it again. ## **Research** ### **Figure, AgiBot and Light Origins report scaling laws, and only AgiBot’s tracks task success** Three companies published data-scaling results in the past two weeks. Each presents its result as evidence that more pretraining data keeps making robots better, but they measure different things. Figure’s [Helix 2.5](https://www.figure.ai/news/helix-2-5-zero-shot-30-home-generalization) reports that pretraining on its human-behavior dataset lifts zero-shot success in 30 unseen homes from 9% to 56%. Its scaling law, though, predicts training loss, not task success. AgiBot’s [GE-Act 2.0](https://www.agibot.com/article/231/detail/120.html) measures success against data. From 300 to 30,000 hours, zero-shot success rises from 17.1% to 44.1%, with no sign of leveling off. Light Origins’ [Light-O1](https://www.lightorigins.com/blog/light-o1) measures only offline prediction loss and pose error. Only AgiBot’s curve shows robots succeeding more often as the data grows. None of the three models is downloadable. ### **GPT-6 Astra scores 22.48% or 35.67% on RoboDojo, depending on the harness** Two papers ran OpenAI’s Astra on the RoboDojo benchmark this week. [RoboProbe](https://arxiv.org/abs/2609.24170) had it pick target positions for the gripper and got 22.48%. [RoboDawn](https://arxiv.org/abs/2609.22966) gave it simple move commands, allowed more of them, and got 35.67%. The two papers also disagree on whether one demonstration helps. Each team's connection of the model to the robot is worth 13 points. ### **An hour of clean data beat twenty hours of volume on a production π0.5 fine-tune** [Dream Machines](https://dream-machines.eu/blog/pi05-fine-tuning) fine-tuned π0.5 on a real unboxing task for a German manufacturer and published the ablations. Adding one hour of clean demonstrations to 21 hours of existing data raised success from 76% to 90%. Changing only how the policy blends action chunks at inference raised the original model from 76% to 93%. Validation loss couldn't distinguish good checkpoints from bad ones. Each test is 40 rollouts, so gaps of a few points are noise. ### **Six open policies sit within 1.58 points on LIBERO and 31 apart on paraphrased instructions** [IndustrialVLA-Bench](https://arxiv.org/abs/2609.25562) ran six public policies, including π0.5 and GR00T-N1.7, under one setup. They land within 1.58 points of each other on standard LIBERO and 31 points apart when the instructions are reworded. [RoboFollow](https://arxiv.org/abs/2609.25636) shows why. When a scene allows only one sensible task, policies ignore the instruction. Standard LIBERO scores no longer tell you which policy to pick. ### **A world model trained on absolute actions collapses on the same motion written as deltas** [Karim and Chlon](https://arxiv.org/abs/2609.23252) trained a world model on absolute positions, then gave it the same motion written as position changes. Goal-reaching success fell from 53% to 15%, even though both formats carry the same information. Nothing raised an error. ### **Protecting 41 more layers takes four-bit π0.5 from 7% to 70.5% success** [VLAQuantBench](https://arxiv.org/abs/2609.25376) tested four-bit quantization across several VLAs. On π0.5, keeping 41 more action-head layers at full precision lifts success from 7.0% to 70.5%. The recipe that helps π0.5 hurts π0, so tune quantization for each model. ### **Eidon AI shut down and put 1,274 hours of egocentric video under CC-BY-4.0** Eidon AI, an egocentric-data company, shut down and released [its whole corpus](https://huggingface.co/datasets/eidon-ai/tracker-pov): 1,274 hours of head-mounted video of household chores, with arm tracking. CC-BY-4.0 means you can train commercial models on it. It comes from only 27 people, which limits its variety. ### **Quick hits** - **[MuJoCo 3.14.0](https://github.com/google-deepmind/mujoco/releases/tag/3.14.0)** — tendon force limits that were silently ignored now apply, so some models will behave differently after upgrading. - **[OpenArm in Isaac Lab](https://github.com/isaac-sim/IsaacLab/issues/7938)** — the hand and TCP links default to 1 kg each, against 0.127 kg and zero in the URDF. - **[Octo distillation](https://arxiv.org/abs/2609.23048)** — an eight-layer student matches the teacher on every offline metric and completes 0 of 72 real trials, against the teacher’s 40. - **[Astronex-World 1.0](https://huggingface.co/Astronex-Lab/Astronex-World)** — a 5B action-conditioned video world model with Apache-2.0 weights, within a point of a 22B model on WBench. - **[InstinctFlash](https://github.com/General-Instinct/InstinctFlash)** — a Jetson Thor serving runtime. Its 33.78x speedup includes running fewer sampling steps, which changes the model’s outputs. - **[LIBERO-VPro](https://arxiv.org/abs/2609.24350)** — a 196,000-episode robustness benchmark with an empty Hugging Face dataset. - **[Skild, Physical Self-Play](https://www.skild.ai/blogs/physical-self-play)** — a humanoid soccer policy trained against itself in simulation, then run on a robot. The post gives no score and names no robot. - **[RoboHarm](https://robocurve.org/roboharm/)** — five harmful instructions, 20 trials each. Claude Fable 5.1 refused 20 of 100, GPT-6 Astra 2 and MolmoAct2 none. ## **Industry** ### **Isaac ROS 5.0 removes NITROS, and zero-copy falls back to CPU without saying so** [Isaac ROS 5.0](https://nvidia-isaac-ros.github.io/v/release-5.0/releases/index.html) removes NITROS, so code that calls it directly needs to be rewritten. The new zero-copy path works only when both nodes share a host, a GPU, and a Linux user. Otherwise, data goes through the CPU with no warning. Check `get_backend_type()=="cuda"` in your launch tests before upgrading. ### **Qualcomm is buying PickNik, and prices a 16 GB robotics board under the Orin Nano Super** Qualcomm agreed to [acquire PickNik](https://www.therobotreport.com/qualcomm-acquires-picknik-robotics-keep-moveit-open-source/), which maintains MoveIt, and says it “intends” to keep MoveIt open source. Its Arduino [VENTUNO Q](https://store.arduino.cc/products/ventuno-q) board is on pre-order at €298.99 with 16 GB of memory. NVIDIA’s 8 GB Orin Nano Super kit costs $399. ### **Mercedes will ship Wayve’s end-to-end driver within two years** Mercedes signed a [production agreement](https://electrek.co/2026/09/22/mercedes-benz-wayve-ai-driver-production-agreement/) to put the Wayve AI Driver in future cars for driver assistance in cities and on highways. It is the third top-10 automaker to sign with Wayve, after Nissan and Stellantis. ### **Also** Alphabet’s Intrinsic released its industrial robot software, [Intrinsic Core](https://www.intrinsic.ai/blog/posts/introducing-intrinsic-core), under Apache-2.0. Boston Dynamics opened a [permanent center](https://bostondynamics.com/news/boston-dynamics-opens-robotics-metaplant-application-center-to-train-humanoid-robots-for-manufacturing-tasks/) at Hyundai’s Georgia plant to collect Atlas training data. D-Robotics, which makes robot chips, [raised $400M](https://www.prnewswire.com/news-releases/d-robotics-completes-400-million-in-series-c-funding-driving-the-robotics-industry-into-a-boom-in-product-categories-302881297.html). ## **Shipped, and Still Owed** - **[ABC-130k](https://huggingface.co/datasets/XDOF/ABC-130k) shows its files again** , still under Apache-2.0. You still have to fill in an access form on Hugging Face to download it. - **Unitree’s 102 G1 datasets carry no license on Hugging Face** , while its[UnifoLM-WLA README](https://github.com/unitreerobotics/unifolm-wla) says they are Apache-2.0. - **The pi0-FAST checkpoints behind last week’s lead are [public](https://github.com/Physical-Intelligence/openpi/issues/1047)** , so you can check the 25-point gap yourself. - **[LIBERO-CTRL](https://github.com/hiroki-oist/LIBERO-ctrl) and [ModAR](https://github.com/adamhung60/ModAR-code) released their code under MIT.** ModAR released no checkpoints. - **[XR-2](https://huggingface.co/datasets/challenge-2026/challenge_data) now has about 705 of its 1,500 claimed hours online.** - **[ZimaBlue](https://github.com/ZimaBlue-WAM/ZimaBlue) removed its plan to release six checkpoints.** ## **Worth Watching** - IROS 2026 runs 27 September to 1 October in Pittsburgh. And we’ll be there too! - Agility holds an analyst and investor day on 6 October. - TrendForce’s September DRAM contract prices are due around 30 September.