Measured model
Measured on a Mac Studio (M5 Max, 36GB)NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
NVIDIA · Ollama name hf.co/ggml-org/NVIDIA-Nemotron-3-Nano-30B-A3B-GGUF:Q4_K_M
On a Mac Studio (M5 Max, 36GB), NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 answers a simple question in about 2 s, gets 77% of the Japanese knowledge questions right, 61% of the math problems, and uses 22.7GB of memory.
These numbers are from before the method changed on Oct 2, 2026 (see how it was measured).
Results
| Measure | Result | Among measured models | Note |
|---|---|---|---|
| Reply time | about 2 s | 15 of 30 | Until one simple common-sense question is answered (median of 50) |
| Knowledge (high school to university) | 77% | 14 of 30 | 200 questions, ±6, 18 cut off (counted wrong) |
| Math (word problems) | 61% | 27 of 30 | 100 questions, ±10, 19 cut off (counted wrong) |
| Memory used | 22.7GB | 30 of 30 | |
| EN→JA translation | 31 pts | 20 of 28 | out of 100 (chrF), about 2.9 s a sentence |
| JA→EN translation | 56 pts | 13 of 28 | out of 100 (chrF), about 3 s a sentence |
"Among measured models" is the place among the 30 models measured on this site (lower is better for time and memory).
Try it on your Mac
- Install the Mac version of Ollama from ollama.com and start it.
- Type this in Terminal. The first time, the model downloads; then you can chat with it.
ollama run hf.co/ggml-org/NVIDIA-Nemotron-3-Nano-30B-A3B-GGUF:Q4_K_M
It used 22.7GB here. To use it comfortably alongside other apps, a Mac with 48GB of memory or more is a good guide (the model then takes about half the memory).
About the model
- Maker
- NVIDIA
- Size
- 31.6B parameters
- Quantization
- Q4_K_M
- Thinks before answering
- Yes
- Longest input at once
- 1,048,576 tokens
- Released (Hugging Face)
- Dec 4, 2025
- Generation speed
- 92.3 tok/s
- Start-up (loading)
- about 1.3 s
- Measured on
- Sep 29, 2026
- Ollama
- 0.34.3
- Hugging Face
- nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
- Series
- Nemotron
Models of similar memory
- DeepSeek-R1-Distill-Qwen-32B (22.4GB of memory, knowledge 81%, reply about 16.8 s)
- qwen3.6:35b-a3b (22.3GB of memory, knowledge 88%, reply about 8.8 s)
- Ornith-1.5-35B-A3B (21.8GB of memory, knowledge 82%, reply about 1.9 s)
- Ornith-1.0-35B (21.6GB of memory, knowledge 88%, reply about 10.1 s)
- glm-4.7-flash:latest (19.5GB of memory, knowledge 77%, reply about 6.3 s)
Compare the main models on the text page. The series page lists every model measured.