{"slug":"3-5-lightning","label":"3.5 Lightning","item_count":3,"day_count":3,"source_count":3,"first_seen":"2026-08-11T13:00:05+00:00","last_updated":"2026-08-19T10:44:16+00:00","generated_at":"2026-09-01T10:05:58.923193+00:00","sources":["aws_ml_blog","nvidia_blog","search_cn_open_weight_labs"],"days":[{"date":"2026-08-11","items":[{"title":"NVIDIA Nemotron 3.5 Lightning and NeMo Switchyard Deliver Faster, Smarter, More Efficient Agentic AI","url":"https://blogs.nvidia.com/blog/nemotron-lightning-switchyard-rtx-dgx","source":"nvidia_blog","type":"news","summary_1line":"As AI shifts from chatbots to autonomous agents, open models are serving market demands for full control over where AI runs and how it’s deployed and evolves. Today, NVIDIA is expanding its Nemotron 3 model family wit...","why_it_matters":"Matches feed focus: agentic.","sid":"9428cc2bc3c996d8","published":"2026-08-11T13:00:05+00:00","editor_note":"NVIDIA's initial announcement: open Nemotron 3.5 Lightning plus a NeMo Switchyard companion release, pitched for agentic workloads."}]},{"date":"2026-08-17","items":[{"title":"NVIDIA Nemotron 3.5 Lightning now available in Amazon SageMaker JumpStart","url":"https://aws.amazon.com/blogs/machine-learning/nvidia-nemotron-3-5-lightning-now-available-in-amazon-sagemaker-jumpstart","source":"aws_ml_blog","type":"news","summary_1line":"NVIDIA Nemotron 3.5 Lightning, an open model built for high-volume agentic workloads, is now available in Amazon SageMaker JumpStart. This post shows how to deploy the 30B Mixture-of-Experts model (3B active), which d...","why_it_matters":"Matches feed focus: agentic.","sid":"0eadff1b96c5b07c","published":"2026-08-17T18:06:33+00:00","editor_note":"First third-party distribution channel — AWS adds the 30B MoE (3B active) build to SageMaker JumpStart, detailing deployment steps."}]},{"date":"2026-08-19","items":[{"title":"The Sequence Frontier Learning - Issue 917: Understanding DeepSeek V4-Pro, GLM-5.3, NVIDIA Nemotron 3.5 Lightning and NeMo Switchyard - TheSequence | Jesus Rodriguez","url":"https://news.google.com/rss/articles/CBMie0FVX3lxTE9iTWREUWM4RGRZUzlKbm5SZTlJQVpDQXNTRjd5dzdYSkZJSk5VY0c4cTEzMEVLWF9RRGFlX0pLV25tTDJqWmJTWTZKMjdBY0tjVGJKMVdqWnYzaEhiVk5rR0tHVWc4OXpLS3dZcVhBX2J0c1hhM1RORkVpUQ?oc=5","source":"search_cn_open_weight_labs","type":"news","summary_1line":"The Sequence Frontier Learning - Issue 917: Understanding DeepSeek V4-Pro, GLM-5.3, NVIDIA Nemotron 3.5 Lightning and NeMo Switchyard TheSequence | Jesus Rodriguez","sid":"7a7406b80e658a40","published":"2026-08-19T10:44:16+00:00","editor_note":"Independent analyst newsletter folds the model into a wider explainer on the week's open-weight frontier releases."}]}],"editorial":{"tldr":"NVIDIA opened its Nemotron 3 family further on Aug 11 with 3.5 Lightning and a companion NeMo Switchyard release, aimed at agentic workloads rather than chatbots. By Aug 17 the 30B-parameter Mixture-of-Experts model (3B active) had its first managed-cloud distribution channel: AWS SageMaker JumpStart.","stale":false,"whats_new":"TheSequence's Aug 19 frontier-model rundown groups Nemotron 3.5 Lightning with DeepSeek V4-Pro and GLM-5.3, two days after AWS added the 30B-MoE (3B-active) build to SageMaker JumpStart.","why_it_matters":"A 30B MoE with only 3B active parameters plus SageMaker JumpStart availability gives platform teams a low-latency, open-weight option for agentic tool-calling without standing up custom serving infrastructure.","take_for_builders":"If you're benchmarking open-weight models for agent tool-calling, try Nemotron 3.5 Lightning via the SageMaker JumpStart path before committing to custom serving — its 3B-active/30B-total split targets that cost/latency profile.","status":{"state":"Available · SageMaker JumpStart","tone":"now","changed":"2026-08-17","detail":"Shipped by NVIDIA on Aug 11 and reached managed-cloud availability via AWS SageMaker JumpStart on Aug 17.","track":[{"label":"announced","detail":"Aug 11","tone":"launch","weight":40},{"label":"SageMaker JumpStart","detail":"Aug 17 → now","tone":"now","weight":60}]},"beats":[{"kicker":"LAUNCH","tone":"launch","headline":"NVIDIA ships Nemotron 3.5 Lightning + NeMo Switchyard for agentic workloads","summary":"Expansion of the Nemotron 3 family, pitched at agent deployments over chat.","sids":["9428cc2bc3c996d8"]},{"kicker":"AVAILABILITY","tone":"now","headline":"30B-MoE (3B active) build lands in AWS SageMaker JumpStart","summary":"First managed-cloud deployment path outside NVIDIA's own channels.","sids":["0eadff1b96c5b07c"]},{"kicker":"COVERAGE","tone":"neutral","headline":"TheSequence groups it with DeepSeek V4-Pro and GLM-5.3 in a frontier-model rundown","sids":["7a7406b80e658a40"]}],"open_questions":["What does NeMo Switchyard actually do — the Aug 11 announcement pairs it with Nemotron 3.5 Lightning but doesn't explain its role.","Will Nemotron 3.5 Lightning reach deployment surfaces beyond SageMaker JumpStart, such as NVIDIA NIM, Hugging Face, or other clouds?"],"provenance":{"0eadff1b96c5b07c":{"status_update":true}},"generated_at":"2026-08-19T15:15:00+00:00"}}