$ cat wiki/models/nemotron-3-5-lightning.md
Nemotron 3.5 Lightning
Spec
| Attribute | Value |
|---|---|
| Developer | NVIDIA |
| Released | 2026-08-11 |
| Announced | 2026-08-11 |
| Context window | 1M |
| Catalogue context window | 262,144 tokens (hosted limit; OpenRouter, 2026-10-22) |
| Pricing | unknown |
| License | OpenMDW-1.1 |
| Availability | Hugging Face (BF16 + NVFP4 open weights), NVIDIA NGC, build.nvidia.com, OpenRouter, DeepInfra, Together AI, Baseten, FriendliAI |
| Catalogue id | nvidia/nemotron-3.5-lightning |
| 30B total / 3B active hybrid Mixture-of-Experts, described as **interleaved | |
| Mamba-2 and MoE layers with select Attention layers** | |
| (source). | |
| It is distilled from NVIDIA's frontier Nemotron 3 Ultra and developed with | |
| the Nemotron Coalition, and ships with multi-token prediction and | |
| DFlash speculative decoding | |
| (source). |
Data cutoffs as reported from the model card: pre-training September 2025, post-training May 2026 (source).
Pricing is unknown rather than free, on the same reading as
Muse Glimmer: the weights are open and no first-party list price was
published in anything read. Third-party serving prices exist and disagree with
each other — DeepInfra at $0.05/M input · $0.20/M output at launch, some
listings at $0.00/$0.00 free tiers, and OpenRouter carrying both a paid and a
:free route
(source).
Those are resellers' prices for open weights, not NVIDIA's, and CLAUDE.md's rule
for a catalogue price applies: any figure inside the provider range is real. The
Catalogue id row is present so the daily spec-check Action can resolve one —
it is copied from the OpenRouter listing URL, not guessed.
Release Date
2026-08-11, weights the same day on Hugging Face
(nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 and the -NVFP4 variant)
and on NVIDIA NGC
(source).
Shipped alongside NeMo Switchyard, an open-source
model-routing library — the two were announced together and the pairing is the
point rather than a coincidence of timing (see ## Compared To).
Benchmarks
NVIDIA's own figures except where the attribution says otherwise. No first-party benchmark table was read — the sandbox could not reach NVIDIA's blog or the model card (source).
| Measure | Figure | Whose number |
|---|---|---|
| Agent productivity (PinchBench) | 86.5% | NVIDIA |
| Output speed | up to 4x vs. comparable dense models | NVIDIA |
| Agentic task completion time | up to 30% faster vs. class | NVIDIA |
| Artificial Analysis Intelligence Index | 24 | Artificial Analysis (third party) |
Two limits. **PinchBench appears in no sources/evals/ snapshot this repo | ||
| holds**, so there is nothing local to check 86.5% against — the same gap recorded | ||
| on Muse Glimmer for MCP Atlas and on Grok Imagine Image 2.0 | ||
| for the image-edit board. And the **Artificial Analysis figure of 24 was read from | ||
| a search extract, not from a snapshot**: this repo's own Artificial Analysis | ||
| captures are produced by the weekly Action, and the most recent | ||
| (source) predates this | ||
| model by nine days, so it carries no row for it. |
The claim shape is worth naming: three of the four figures are about speed or cost, not capability. That is consistent with a model positioned as the cheap step inside someone else's workflow rather than as a frontier competitor.
Use Cases
NVIDIA's stated framing is explicitly subordinate: a frontier reasoning model — Nemotron 3 Ultra, or GPT-5.6, which NVIDIA names — plans and orchestrates a workflow, while a smaller specialised model such as this one performs targeted steps: code review, tool use, security-alert monitoring, answering billing questions (source).
The 1M context window with 3B active parameters is the combination that makes that role affordable: long-horizon agent state can stay in context without paying dense-30B compute per token.
Compared To
| Model | Params | Active | Context | Licence | Weights |
|---|---|---|---|---|---|
| Nemotron 3.5 Lightning | 30B MoE | 3B | 1M | OpenMDW-1.1 | released |
| Muse Glimmer | 30B dense | 30B | 131,072 | Apache 2.0 | released |
| Laguna S 2.1 | 118B MoE | 8B | 1M | OpenMDW-1.1 | released |
| Gemma 4 12B | 12B dense | 12B | unknown | unknown | released |
| The first two rows landed one day apart at the same parameter count, and they | |||||
| are not the same product. Muse Glimmer is dense, so all 30B activate, and its | |||||
| pitch is that it fits on one consumer GPU at 4-bit. Lightning is sparse, so | |||||
| 3B activate, and its pitch is throughput at long context inside a server-side | |||||
| agent loop. Same headline number, different constraint being solved — a reminder | |||||
| that "30B" has stopped being a useful comparison unit. |
The licence is the other difference, and it is not new to this wiki: OpenMDW-1.1 is the same licence Laguna S 2.1 shipped under on 2026-07-21, where it is recorded as permitting use, modification and redistribution including commercial use (source). Coverage of this release describes it the same way — open for commercial use without material restrictions (source). Two unrelated vendors picking the same non-Apache open-model licence within three weeks is the more interesting fact, and worth watching rather than concluding from.
Conflicting Reports
The announced 1M capacity is distinct from a serving default. On 2026-10-22 NVIDIA's published BF16 configuration sets max_position_embeddings to 262,144 tokens, matching the OpenRouter hosted limit. The announced capacity is retained; it does not promise that every hosted endpoint accepts 1M tokens.
Separately, the Switchyard cost claim is reported two ways — NVIDIA's "nearly one-third of Opus 4.8 alone" and VentureBeat's headline "cutting task costs to a third in its own tests" — which are the same claim, and both label it as NVIDIA's internal benchmark (source).
Open Questions
- Training compute and token budget — undisclosed. The distillation lineage from Nemotron 3 Ultra is stated; nothing else about the training run is.
- What PinchBench is, and who maintains it. An 86.5% on a suite this repo has no snapshot of is a number with no denominator.
- Does 1M context hold up? No long-context retrieval measurement surfaced in anything read — the figure is an architectural capacity, not a demonstrated one.
- The OpenMDW-1.1 terms. Named, characterised, not read.
- Was the evaluation harness released with the weights? The distinction Eval Harness Configuration exists to record, unresolved here.
Sources
- NVIDIA Blog — Nemotron 3.5 Lightning and NeMo Switchyard Deliver Faster, Smarter, More Efficient Agentic AI (source)
- NVIDIA Technical Blog — Nemotron 3.5 Lightning Delivers Fast, Accurate Specialized Task Execution for Long-Running Agents
- Hugging Face — nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16
- Hugging Face — nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4
- DeepInfra — Nemotron 3.5 Lightning day-zero access
- Models.dev — Nemotron 3.5 Lightning 30B A3B pricing, providers, specs
- explainX — Nemotron 3.5 Lightning: 30B Open MoE That Runs on 1 H100