NVIDIA · OpenMDW 1.1
Nemotron 3.5 Lightning
NVIDIA’s August 2026 30B-A3B, released with its training data under a permissive licence.
How much memory it needs
The model file, plus memory for the conversation, plus about 1 GB for the app. Longer conversations need more.
| Size | Version | File | At 8K | At 32K | At 128K |
|---|---|---|---|---|---|
| 32B (3B active) | Q4_K_M | 25.5 GB | 26.5 GB | 26.7 GB | 27.3 GB |
| Q8_0 | 35.0 GB | 36.0 GB | 36.2 GB | 36.8 GB |
Which computers run it
The best size for each machine at an 8K-token conversation. Speeds are estimates from memory bandwidth.
| Computer | Best version | Fit | Speed |
|---|---|---|---|
| Laptop or PC, no graphics card, 32 GB | 32B Q4_K_M | Too big | — |
| RTX 4060 · 8 GB | 32B Q4_K_M | Spills into system RAM | 15–30tok/sSmooth |
| RTX 3060 · 12 GB | 32B Q4_K_M | Spills into system RAM | 18–35tok/sSmooth |
| RTX 5060 Ti · 16 GB | 32B Q4_K_M | Spills into system RAM | 25–45tok/sSmooth |
| RTX 4090 · 24 GB | 32B Q4_K_M | Spills into system RAM | 60–110tok/sFast |
| RTX 5090 · 32 GB | 32B Q4_K_M | Runs well | 130–260tok/sFast |
| Mac with M4 · 16 GB | 32B Q4_K_M | Too big | — |
| Mac with M4 Pro · 48 GB | 32B Q8_0 | Runs well | 30–65tok/sFast |
| Mac with M5 Max · 128 GB | 32B Q8_0 | Runs great | 50–110tok/sFast |
| Ryzen AI Max+ 395 · 128 GB | 32B Q8_0 | Runs great | 25–50tok/sFast |
| DGX Spark · 128 GB | 32B Q8_0 | Runs great | 30–55tok/sFast |
| RTX PRO 6000 · 96 GB | 32B Q8_0 | Runs great | 110–220tok/sFast |
| Mac Studio with M5 Ultra · 256 GB | 32B Q8_0 | Runs great | 50–110tok/sFast |
| Mac Studio with M3 Ultra · 512 GB | 32B Q8_0 | Runs great | 40–85tok/sFast |
What it is good and bad at
Good at
- Tiny conversation memory
- Open data and recipes
Not good at
- About 25 GB at Q4: needs a 32 GB card or a Mac
Start it
Install Ollama, then run one of these in a terminal. LM Studio users can search the same name in the app.
ollama run nemotron-3.5-lightning:30b-a3b