Nemotron 3.5 Lightning
Intelligence benchmarks
Artificial Analysis indexes - compared with the best open and proprietary models
Intelligence
5.5
AA Index
Coding
0.2
AA Index
Agentic
1.9
AA Index
Intelligence Index - Nemotron 3.5 Lightning vs. the field
Best open-weight models (you can run locally) and leading proprietary models for context.
Coding Index comparison
Agentic Index comparison
Benchmark data from Artificial Analysis · updated 2026-11-12.
Hosted API cost & speed
What it costs to rent this model instead of running it locally
| Metric | Value |
|---|---|
| Input (per 1M tokens) | $0.06 |
| Output (per 1M tokens) | $0.20 |
| Median speed (hosted API) | 271 tok/s |
Median across hosted providers, measured by Artificial Analysis - updated 2026-11-12. Local llama.cpp speed depends on your hardware — see the hardware pages for estimates.
Standard benchmarks
Performance across standard evaluations
| Benchmark | Score |
|---|---|
| GPQA Diamond | 74.3 |
| HLE | 10.6 |
| SciCode | 32.1 |
Source: Artificial Analysis component evals - updated 2026-11-12.
Will it run on your hardware?
Pick your GPU memory - see which quantizations fit, and the cheapest card for the rest
Need an exact figure for your context length? Use the VRAM calculator.
Run it locally
Copy-paste - running in under a minute
vllm serve nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4New to this? Start with Ollama · serve to many users with vLLM.
Deep dive
Notes, sources, and the full write-up
Nemotron 3.5 Lightning is a 17.8B-parameter other model from NVIDIA. It scores 5.5 on the Artificial Analysis Intelligence Index (coding 0.2). At Q4_K_M it needs roughly 10 GB of VRAM, placing it in the 8–12 GB GPU hardware tier.
BenchmarksLink to this section
Artificial Analysis Intelligence Index - Nemotron 3.5 Lightning vs. leading closed models:
| Model | Intelligence | Coding | GPQA |
|---|---|---|---|
| Nemotron 3.5 Lightning | 5.5 | 0.2 | 74.3 |
| Claude Fable 5.2 | 64.6 | 53.1 | - |
| Gemini 4 Carbon | 60 | 50.4 | - |
| Claude Opus 5.5 | 59.1 | 51.7 | - |
| GPT-6 Astra | 56.9 | 49.1 | 96.1 |
| GPT-6.1 Sol | 54.6 | 46.8 | - |
Source: Artificial Analysis (2026-11-12).
PopularityLink to this section
Nemotron 3.5 Lightning has 19,250 downloads in the last month on HuggingFace and 205 likes.
From the community
What users are saying
“The snippet was passing --speculativeconfig.method dspark without pointing to the DSpark draft model, which the other spec decoding snippets pass correctly. Now the snippet specifies the right thing [the method is correctly inferred from the spec model as dspark, in any case], just set $DSPARKCKPT as mentioned above.”
“NVIDIA Nemotron 3.5 Lightning NVFP4 on RTX 5090: vLLM 0.27.1 Marlin failure Proposed Hugging Face discussion for [nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4](https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4).”
“Summary Benchmarked speculative decoding variants — baseline vs. DSpark and DFlash, — for nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4 on a GX10 (DGX Spark, GB10, 121GB unified memory), across chat (ShareGPT) and code (HumanEval) workloads.”
Frequently asked
Quick answers to common questions
How much VRAM does Nemotron 3.5 Lightning need?
Nemotron 3.5 Lightning with 17.8B parameters needs approximately 10 GB at Q4_K_M quantization. Use our VRAM calculator for an exact estimate.
Is Nemotron 3.5 Lightning better than other nvidia models?
Nemotron 3.5 Lightning has 17.8B parameters with 1,048,576 context - a strong choice for long-context, general-chat.
What license is Nemotron 3.5 Lightning under?
Nemotron 3.5 Lightning is released under the other license, making it suitable for most commercial and personal projects.
What hardware runs Nemotron 3.5 Lightning well?
With 17.8B parameters, Nemotron 3.5 Lightning requires adequate VRAM. High-end GPUs like the RTX 4090 (24GB), RTX 5090 (32GB), or Mac Studio with unified memory are good options. Check our hardware directory for specific recommendations.
What is the best quantization for Nemotron 3.5 Lightning?
Q4_K_M is the recommended sweet spot - ~98% of FP16 quality at ~27% of the size. Q5_K_M (~13 GB) is an option if you have spare VRAM. Use our VRAM calculator to compare.
How long can Nemotron 3.5 Lightning's context window handle?
Nemotron 3.5 Lightning supports a 1,048,576-token context window - enough for very long documents, codebases, or multi-turn conversations. Real-world usable context may vary by implementation.
What models compete with Nemotron 3.5 Lightning?
Nemotron 3.5 Lightning competes with other models in its class. Browse our model directory for comparisons, benchmarks, and community reviews to find the best fit.
Can my GPU run Nemotron 3.5 Lightning?
Fit and estimated speed on popular hardware
Nearby options
Similar models and compatible hardware by spec