#NVIDIA just released a whole ecosystem for human(oid) motion and robot learning from human data. 🚀🦾 Data, as we all know, is the key to scaling AI models. To accelerate the field of Embodied AI, we have open-sourced a full stack of models and tools to capture, generate, retarget, and simulate human(oid) motion data at scale, along with a massive high-quality dataset and a standard human skeletal representation, SOMA, to make them all seamlessly communicate with each other. The entire suite is available under the Apache 2.0 license. 1️⃣ SOMA: A universal interface to unify all parametric human body models (SOMA-shape, SMPL, MHR, etc.) into a standard skeletal representation, eliminating the need for custom adapters or model-specific retargeting. 🔗 lnkd.in/gsxhiJnn 2️⃣ Kimodo: High-fidelity, controllable text-to-motion generation for both humans and humanoid robots. 🔗 lnkd.in/gCc84XnX 3️⃣ GEM: A global human pose estimation method from in-the-wild videos, natively compatible with SOMA. 🔗 lnkd.in/g_QAvRjn 4️⃣ Bones-SEED: A massive dataset of 150k+ motions in SOMA format, including data already retargeted for the Unitree G1, created with our partners at Bones Studio. 🔗 lnkd.in/gfx-QD-w 🔗 lnkd.in/gyNdTwQx 5️⃣ SOMA Retargeter: A dedicated tool for seamless motion retargeting from the SOMA skeleton to the Unitree G1. 🔗 lnkd.in/gqz9Na-H 6️⃣ ProtoMotions: Our high-performance simulation framework for training digital human(oid)s via RL, now with native SOMA support. 🔗 lnkd.in/gmvMikMU This is just the beginning, and we have much more in the pipeline. Excited to see what the community builds next! #NVIDIA #GTC #GTC2026 #Robotics #EmbodiedAI #PhysicalAI @NVIDIAAI

one day you wrote console.log("here") for the last time and didn't even know it

We have never seen a model do this before, but GLM-5.3 sometimes reasons (unprompted) about how its current actions could become training data and impact its future self. 🧵

Very excited to announce our new models: Nemotron 3.5 Super and Lightning! This also feels like a good time to share that a few months ago I joined the amazing Nemotron Pretraining team. We're cooking some very large and very strong models for you! Here is a tech summary for 3.5: * Super is a 120B total / 12B active vision language model (VLM) * Lightning is a 30B total / 3B active text-only LLM * Super scores 51.78% on SWE-bench Pro (better than Nemotron 3 Ultra despite ~4.6× fewer active params!) * arch: MoE Hybrid Mamba-Transformer designed for efficient inference, with a 1M token ctx window, and native speculative decoding (via MTP) * training is done across our GB200 and GB300 (and deployment validated across all imaginable SKUs, including consumer (RTX) and DGX Spark) A few details on how we train these: -- Pretraining * pipeline: initialize with Nemotron 3 base -> custom MTP initialization + 100B-token MTP warm-up -> 1T-token distillation from the next larger N3 teacher * Lightning then goes through a 21B token long-context (LC) stage (@ 1M ctx) * Super goes through: - Indic CPT: 100B tokens at 8K (10B Indic + 90B replay, we built a synthetic corpus with 2.83T toks across 22 Indian langs! Indic benchmarks improve by up to 9.82%) - LC extension: 35B tokens at 1M - Agentic SWE mid-training: 65.4B tokens at 256K (uses whole repos, issue-PR pairs, commit histories, etc.); DeepSWE pass@1 improves from 6.4 -> 10.4 - LC recovery: 9B tokens at 1M (as we observed regression after mid-training) -- Post-training * Super: two-stage SFT -> joint RLVR/RLHF -> specialist teachers -> two-stage MOPD -> MTP boosting * Lightning uses a lighter text-only pipeline: single-stage SFT -> RLVR * Super's SFT transforms the text-only base into a VLM using vision, agentic, reasoning, chat, safety, LC data * A few examples of RLVR environments: - Olympiad-level competitive programming - Finance agents using SEC search, document parsing, evidence retrieval, and tool use * Specialist teachers include Agentic, Reasoning, Video reasoning and tracking, Vision tool-use, Vision-reasoning, and an Alignment/RLHF teacher * Two-stage MOPD consolidates these specialist capabilities into one model * Post-training ends with MTP boosting to better align the drafter and reduce recursive train–inference mismatch; this improves Super's SPEED-Bench acceptance length by 21.8% (to 4.612 tokens) -- Inference The model is highly optimized for inference (with plenty of quantized options to choose from)! A few data points: * On a GB200, Super reaches 4,856 output toks/s/GPU with ≥50 toks/s/user on a 20-turn agentic workload * Super provides 2.34x the throughput of Qwen3.5 122B-A10B and 1.90x the throughput of Qwen3.8 Flash-Next * Native MTP is complemented by standalone parallel drafters: LiLiCorr for Super, and DFlash and DSpark for Lightning Looking forward to seeing what you do with 3.5! Phase shift for Nemotron incoming! :)) (with the usual this will be our weakest model ever, but really really so) paper: research.nvidia.com/labs/nem…