Open Source Isaac 0.5: 36B Embodied Model Trained on Web Videos
lukas_m_ziegler · x · 2026-08-27
Perceptron Inc (ex-Meta Chameleon team) released Isaac 0.5, an open-source embodied foundation model that turns internet video into robot training data. The same weights handle video QA, object pointing, tracking, task reporting, and action generation.
Key specs:
- Scale: 36B-parameter sparse model, trained on 35+ robot systems and 3T multimodal tokens.
- Performance: Scores 62.6 on ScreenSpot-Pro grounding vs 54.8 for the strongest Qwen3-VL.
- Efficiency: Achieves this at 8.5× lower inference cost (26.9 TFLOP vs 228.4).
- Tech: Uses "Null Experts" where each token selects how many experts to activate, allocating compute dynamically.
Related event: Perceptron Releases Isaac 0.5, a 36B Open-Weight Embodied Foundation Model(5 posts)→
More from Embodied
- Building robotic textile production lines for Earth and the lunar economy — MatthewChang · 2026-08-27
- SF Startup Lightberry Launches Lumi, a Custom Unitree G1 Humanoid — Distinct-Question-16 · 2026-08-27
- Robotics Value Accrual: Embodied AGI and Deployment Layer — chris_j_paxton · 2026-08-27
- Actuate 2026 recap: Data as infrastructure, real-world deployment as the test — Sethwinterroth · 2026-08-27
- NSF invests $90M in three new Science and Technology Centers, including human-robot co-adaptation — yudapearl · 2026-08-27
- Sai-Fi: Connects Meta Glasses to Autonomous Computer Agent — xwang_lk · 2026-08-27