QuantDiff

Best Nemotron 3.5 Bolt E4B quant for a 24GB GPU

Short answer: UD-Q8_K_XL from unsloth . Of the 8 Nemotron 3.5 Bolt E4B files we tested that fit fully in 24GB with a 8192-token context, it has the smallest diff from the BF16 model (mean KLD 0.0009, same top token 98.7%).

Smallest diff that fits 24GB
UD-Q8_K_XL by unsloth
File size
7.63 GB
(mean)
0.0009
98.7%
88.0%
Accuracy
Generation ()
31 t/s on Apple M2 Pro 32GB full GPU
NVIDIA-Nemotron-3.5-Bolt-E4B-UD-Q8_K_XL.gguf
How to run this file

From unsloth/NVIDIA-Nemotron-3.5-Bolt-E4B-GGUF, file NVIDIA-Nemotron-3.5-Bolt-E4B-UD-Q8_K_XL.gguf.

(downloads it, then starts a chat; type /bye to quit):

ollama run hf.co/unsloth/NVIDIA-Nemotron-3.5-Bolt-E4B-GGUF:NVIDIA-Nemotron-3.5-Bolt-E4B-UD-Q8_K_XL.gguf

: open the Discover tab, search unsloth/NVIDIA-Nemotron-3.5-Bolt-E4B-GGUF, pick the UD-Q8_K_XL download, then load it from the Chat tab.

Step-by-step guide

Runner-up
UD-Q6_K_XL by unsloth
File size
6.62 GB
(mean)
0.0060
96.7%
89.0%
Accuracy
Generation ()
36 t/s on Apple M2 Pro 32GB full GPU
NVIDIA-Nemotron-3.5-Bolt-E4B-UD-Q6_K_XL.gguf
How to run this file

From unsloth/NVIDIA-Nemotron-3.5-Bolt-E4B-GGUF, file NVIDIA-Nemotron-3.5-Bolt-E4B-UD-Q6_K_XL.gguf.

(downloads it, then starts a chat; type /bye to quit):

ollama run hf.co/unsloth/NVIDIA-Nemotron-3.5-Bolt-E4B-GGUF:NVIDIA-Nemotron-3.5-Bolt-E4B-UD-Q6_K_XL.gguf

: open the Discover tab, search unsloth/NVIDIA-Nemotron-3.5-Bolt-E4B-GGUF, pick the UD-Q6_K_XL download, then load it from the Chat tab.

Step-by-step guide

Fastest that fits
UD-Q3_K_XL by unsloth
File size
3.93 GB
(mean)
0.0510
91.2%
88.5%
Accuracy
Generation ()
45 t/s on Apple M2 Pro 32GB full GPU
NVIDIA-Nemotron-3.5-Bolt-E4B-UD-Q3_K_XL.gguf
How to run this file

From unsloth/NVIDIA-Nemotron-3.5-Bolt-E4B-GGUF, file NVIDIA-Nemotron-3.5-Bolt-E4B-UD-Q3_K_XL.gguf.

(downloads it, then starts a chat; type /bye to quit):

ollama run hf.co/unsloth/NVIDIA-Nemotron-3.5-Bolt-E4B-GGUF:NVIDIA-Nemotron-3.5-Bolt-E4B-UD-Q3_K_XL.gguf

: open the Discover tab, search unsloth/NVIDIA-Nemotron-3.5-Bolt-E4B-GGUF, pick the UD-Q3_K_XL download, then load it from the Chat tab.

Step-by-step guide

Other budgets: 8GB12GB16GB · all 8 Nemotron 3.5 Bolt E4B quants

How we picked

We estimate VRAM need as file size + f16 KV cache for 8192 tokens + about 0.6GB of compute buffers, and keep 0.5GB free for your desktop. Among files that fit, we rank by mean KL divergence against the BF16 model's logits, then by tool-call score. If you need a longer context, step down one quant level or quantize the KV cache (-ctk q8_0 -ctv q8_0). Methodology.

GPU too small? Rent one.

Run a 32B model at Q6/Q8, or a 70B, on a cloud GPU for an hour or two instead of buying hardware.

RunPod

Cloud GPU pods with 24-80GB cards (RTX 4090 / 5090 / A100 / H100) for big quants or BF16 reference runs.

Rent on RunPod
Vast.ai

GPU marketplace for renting a 24-48GB card for an evening of testing.

Rent on Vast.ai
Your GPU cloud here

Sponsor slot. Sponsors never influence scores or rankings.

Sponsor QuantDiff

Referral links: QuantDiff may earn credit if you sign up. Referral partners have no say in scores or rankings.