NVIDIA-Nemotron-3.5-Bolt-E4B-4GB · Q4_K_M
This page aggregates 1 real-world runs of NVIDIA-Nemotron-3.5-Bolt-E4B-4GB (Q4_K_M) on NVIDIA H20 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
61.75 tok/s
Decode
Decode speed
319.38 tok/s
Prefill
Prefill speed
1.57 s
TTFT
Time to first token
15.88 GB
VRAM
VRAM usage
1
Measured runs
1
Independent sources
Other
Source platforms
12 days ago
Last verified
Performance
- llama.cpp · Q4_K_M (current)Decode 61.75 · Prefill 319.38 ·
- Strata · Q4_K_M Decode 91.09 · Prefill 326.32 ·
Core figures
- Decode (avg)
- 61.75 tok/s
- Prefill (avg)
- 319.38 tok/s
- TTFT (avg)
- 1.57 s
- VRAM (avg)
- 15.88 GB
- MTP acceptance rate
- 0.55
- TTFB
- 0.49 GB
- Power draw
- 366.4 W
Configuration
Member-level fields are taken from the most recent run
- Model
- NVIDIA-Nemotron-3.5-Bolt-E4B-4GB
- Quantization
- Q4_K_M
- Framework
- llama.cpp
- Version
- v0.44
- Context length
- 65536 tokens
- Batch size
- 256
- GPU layers
- 99
- Flash Attention
- On
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- NVIDIA H20
- Nominal VRAM
- 97 GB
- Measured VRAM (avg)
- 15.88 GB
- OS
- Ubuntu 22.04
- Driver
- 565.57
- CUDA
- 12.4
- Power draw
- 366.4 W
Get started
Original model resources
Base model resources may not match this quantization. Use the verified deployment weights above when available.
Sources & evidence
1 measured records in total, each traceable to its original source
v0.44 · Ubuntu 22.04 · CUDA 12.4 · 65536 ctx
61.75 tok/s
Decode
319.38 tok/s
Prefill
1.57 s
TTFT
15.88 GB
VRAM
0.55
MTP
366.4 W
Power draw
L2 batch seed 2026-10-24: llama.cpp on hw_id=56 running model_id=97 at Q4_K_M