NVIDIA-Nemotron-Labs-Dory1.0-4B: a 4B looped (latent-recurrent) normalized-GPT — 8 input layers + a shared 20-layer recurrent block (run 1/2/4 times) + 8 output layers, 256K context, OpenMDW-1.1 license (MIT-like); shipped quietly to Hugging Face on 2026-10-16 with no official blog post. SWE-bench Verified at 1/2/4 loops: 50.3%/57.4%/73.4% (the 4-loop figure is a cited score, not fully NVIDIA-measured), with the 2-loop setting roughly matching similarly sized dense models Qwen3.5-4B/3.5-9B. The pitch is that one checkpoint can trade inference compute for accuracy at serving time by changing the loop count, no retraining or model swap needed; vLLM/llama.cpp support is still under PR review, and there's no official or third-party API pricing yet.
Nemotron covers text reasoning, multimodal understanding, and safety classification. Ultra, Super, Lightning, Nano Omni, and Content Safety differ in architecture, context, and licensing; model specifications and API limits must be checked separately.
TensorRT-LLM is NVIDIA's open-source LLM inference library (Apache 2.0). It offline-compiles model weights and compute graphs into optimized TensorRT engines, then serves them with custom CUDA kernels, in-flight batching, and multi-dimensional parallelism. The cost: NVIDIA GPUs only, compilation takes tens of minutes, and switching models or quantization means rebuilding.
Triton Inference Server serves TensorRT, ONNX, PyTorch, and other models through consistent HTTP and gRPC APIs. Its defining tools are the model repository, dynamic batching, instance groups, and ensembles—not LLM-specific KV-cache scheduling.
NVIDIA's generative AI line has four exams: NCA-GENL and NCA-GENM ($125 each, associate), NCP-GENL and NCP-AAI ($200 each, professional). Three decision inputs no other vendor forces on you. One: both professional exams still show 'Coming soon' next to Register, so any near-term plan is down to the two associates. Two: the cost of the official recommended training differs sharply by level. In autumn 2026 the two associate exams switched to one free open-source set of materials plus one course, while the two professional exams still list five paid courses, $1,620 at list price for NCP-GENL. Three: the official web pages and PDFs differ in a few places — two NCP-AAI weights do not match, and two cells of NCP-GENL's web table carry misplaced text — so for those two exams follow the PDF or budget a range. Lock-in also varies sharply: NCP-AAI is 7% NVIDIA-specific, NCP-GENL is 31% GPU and model-compression work.
NCA-GENL is usually what a job posting means by 'NVIDIA Generative AI / LLM certification.' But the official blueprint diverges sharply from the name — Core Machine Learning and AI Knowledge 30%, Software Development 24%, Experimentation 22%, Data Analysis 14%, Trustworthy AI 10% — with LLM and RAG content scattered at bullet level rather than forming a domain, alongside spaCy, NumPy, Keras, and cross validation. The other thing to know first: the official recommended training was replaced in autumn 2026, and each area now lists just two items, free open-source deep learning course materials on GitHub and one LLM application development course. Official specs: $125, 1 hour, 50–60 items, two-year validity, English only, pass/fail with no score reported.
NCA-GENM matches NCA-GENL on price, length, and level but not on emphasis: Experimentation rises to 25% (the heaviest), Core ML drops from 30% to 20%, and two new areas appear — Multimodal Data 15% and Performance Optimization 10%. The content covers U-Net, CLIP, diffusion models, multimodal loss functions, attention maps, and NVIDIA's Riva / NeMo / Triton / ACE SDKs. The official recommended training was replaced in autumn 2026: it is now one free open-source set of deep learning materials on GitHub plus one multimodal agent course, assigned differently per area, with the heaviest area, Experimentation, getting only the free materials. Official specs: $125, 1 hour, 50–60 items, two-year validity, English only.
NCP-AAI is NVIDIA's professional-level agentic AI credential — $200, 120 minutes, 60–70 items, two-year validity. Of its ten areas the heaviest are Agent Architecture and Design at 15% and Agent Development at 15%; only NVIDIA Platform Implementation at 7% is explicitly tied to NVIDIA products, and the rest is general agentic engineering. Two things to know first: registration is not open (the Register button carries a 'Coming soon' label), and the web page and the PDF study guide differ on two weights (Deployment and Scaling 13% versus 5%, Run/Monitor/Maintain 5% versus 7%), so budget those as a range.
NCP-GENL is NVIDIA's professional-level LLM credential — $200, 120 minutes, 60–70 items. What separates it from every other GenAI exam is where the weight sits: Model Optimization 17% plus GPU Acceleration 14% is 31% on quantization, distillation, pruning, distributed parallelism, and CUDA profiling — not on calling APIs. Two things to know first: the Register button says Coming soon, so you cannot sit it yet; and the description text for Model Optimization and Fine-Tuning in the web page's weight table is misplaced, so prepare those two areas from the official PDF study guide.
The NVIDIA DGX Spark is powered by the GB10 Grace Blackwell Superchip, 128 GB of unified memory, and delivers 1 petaFLOP of FP4 compute — starting at around $3,999 USD. It lets developers run 200B-parameter models locally and fine-tune 70B models, making it the most accessible NVIDIA AI development platform available today.