# LLM-Benchmark Konfiguration # Modell : Nemotron-H-47B-Reasoning-128K # Engine : llama.cpp # Run-ID : run-20260817-174452-cfa36c # GPU : 4x AMD Radeon AI PRO R9700 # CPU : AMD Ryzen Threadripper PRO 7955WX 16-Cores # RAM : 184 GB bench@llm-benchmark:~$ /home/godcore/llama.cpp/build/bin/llama-server \ -m /home/godcore/.cache/huggingface/hub/models--bartowski--nvidia_Nemotron-H-47B-Reasoning-128K-GGUF/snapshots/ce89f02cd0eafe3931acec38f4d91a654ae1ca4d/nvidia_Nemotron-H-47B-Reasoning-128K-Q4_K_M.gguf \ --alias nvidia_Nemotron-H-47B-Reasoning-128K \ -ngl 999 \ -fa on \ -sm layer \ -c 49152 \ -np 12 \ --jinja \ --host 0.0.0.0 \ --port 8000

Engine?Die Inferenz-Software, die das Modell ausliefert (z.B. vLLM oder llama.cpp). Sie bestimmt Geschwindigkeit, unterstuetzte Modellformate und welche Parameter ueberhaupt verfuegbar sind.	llamacpp
Modellalias?Der Name, unter dem das Modell ueber die API angesprochen wird. Genau dieser Wert muss im Request-Feld 'model' stehen.	nvidia_Nemotron-H-47B-Reasoning-128K
Kontextlaenge?Maximale Anzahl Tokens (Eingabe + erzeugte Ausgabe zusammen), die das Modell pro Anfrage verarbeiten kann.	49152
Modellpfad?Pfad zur GGUF-Modelldatei, die geladen und ausgeliefert wird.	/home/godcore/.cache/huggingface/hub/models--bartowski--nvidia_Nemotron-H-47B-Reasoning-128K-GGUF/snapshots/ce89f02cd0eafe3931acec38f4d91a654ae1ca4d/nvidia_Nemotron-H-47B-Reasoning-128K-Q4_K_M.gguf
GPU-Layer?Anzahl der auf die GPU ausgelagerten Modell-Layer. Hoeher = mehr VRAM und schneller; der Rest laeuft auf der CPU. 999 = alles auf GPU.	999
fa	on
Split-Mode?Verteilung ueber mehrere GPUs: none (nur eine GPU), layer (Layer aufteilen) oder row (Tensoren zeilenweise).	layer
Kontext?Groesse des Kontextfensters in Token. 0 = der beim Training verwendete Kontext des Modells.	49152
np	12

All benchmarks of this model To leaderboard

Model comparison

Nemotron-H-47B-Reasoning-128K on various hardware

All published performance runs of this model – each bubble a variant: position = prefill (X) × generation (Y), bubble size = number of runs. Closer to the top right = faster. ★ Marked gold = this benchmark.

GPUby graphics card

NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 203,5 tok/sNVIDIA GeForce RTX 5090 197,2 tok/sNVIDIA RTX PRO 6000 Blackwell Workstation Edition 174,9 tok/s★ AMD Radeon AI PRO R9700 28,1 tok/s this runNVIDIA GeForce RTX 5070 Ti 20,4 tok/sNVIDIA GeForce RTX 3090 Ti 16,9 tok/sAMD Radeon 8060S Graphics 10,7 tok/s

CPUby processor

AMD Ryzen Threadripper PRO 9965WX 24-Cores 203,5 tok/sAMD Ryzen 7 5800X3D 8-Core Processor 197,2 tok/sAMD Ryzen 9 9950X 16-Core Processor 174,9 tok/s★ AMD Ryzen Threadripper PRO 7955WX 16-Cores 28,1 tok/s this runAMD Ryzen Threadripper PRO 5975WX 32-Cores 20,4 tok/sAMD Ryzen 9 8945HX with Radeon Graphics 16,9 tok/sAMD RYZEN AI MAX+ 395 w/ Radeon 8060S 10,7 tok/s

MBby mainboard

★ ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE 203,5 tok/s this runASUSTeK COMPUTER INC. ROG STRIX B550-A GAMING 197,2 tok/sASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI 174,9 tok/sASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 20,4 tok/sMeigao Innovation Technology (Shen Zhen) Co., Ltd DRFXL (MotherBoard Series) 16,9 tok/sBosgame AXB35-02 (BeyondMax Series) 10,7 tok/s

ENGby engine

★ llama.cpp 203,5 tok/s this run

DRVby driver

unbekannt 203,5 tok/s

💰 Economics

Economics of this run

Operating cost, TCO and comparison with the next-best runs of the same model at identical concurrency (5× concurrent). Methodology →

⚙️ ConfigurationAll metrics and charts below follow these settings – based on a 24-month runtime.Save to URL Reset

⚡ Electricity price EUR/kWh

⚙️ System utilization 100 %

🖥️ Acquisition EUR

🔌 Idle 145 W

⚡ TDP 1271 W

☁️ External LLM (API) show

Input €/1M Output €/1M Input per output token

Electricity0.30 EUR/kWh

Avg power (incl. idle)1,271 W estimated (TDP)GPU 1,200 + CPU 46 + Board 25 W full load

Avg cost / hourEUR 0.38

Electricity / 1M tokensEUR 4.23

Token / kWh70.92K

Acquisition (system)EUR 11,114 full priceGPU EUR 5,600 · CPU EUR 1,399 · Board EUR 1,299 · RAM EUR 2,576 · PSU EUR 240

Electricity (2 years)–

TCO (2 years)EUR 17,794

Output tokens (2 years)1.58B

☁️ External LLM (API) – comparison

External LLM cost (2 years)–

Savings vs. external (2 years)–

All values above and the charts below take the configured system utilization into account: at X% the system generates only X% of the time, the rest it idles (145 W). Cost per hour drops (more idle), cost per token rises.

Cost over 2 years – electricity only

Cost over 2 years – incl. acquisition (TCO)

Speed vs. tokens per euro

Euro per 1M tokens

Comparison vs. API – economics per benchmark

	Nemotron-H-47B-Reasoning-128K4x AMD Radeon AI PRO R9700	Nemotron-H-47B-Reasoning-128KNVIDIA RTX PRO 6000 Blackwell Workstation Edition	Nemotron-H-47B-Reasoning-128KNVIDIA GeForce RTX 5090	Nemotron-H-47B-Reasoning-128K3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
Electricity cost (24 mo.)	–	–	–	–
Acquisition cost	–	–	–	–
Total cost (TCO)	–	–	–	–
Generated tokens (24 mo.)	–	–	–	–
Token price via API	–	–	–	–
Break-even point (days)	–	–	–	–
Result (savings / extra cost)	–	–	–	–

4× AMD Radeon AI PRO R9700 · this run 25 tok/s NVIDIA RTX PRO 6000 Blackwell Workstation Edition 128 tok/s NVIDIA GeForce RTX 5090 120 tok/s 3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 118 tok/s

Comparison with up to 3 next-best runs of this model at the same concurrency (at least one on different hardware). Power = GPU TDP + CPU (idle + 15 %) + board (estimated), acquisition = full system (GPU + CPU + board + RAM + PSU), prices = stored market prices.

Contributed by

Mario Alka Administrator

@marioalka

Ich bin Unternehmer, Softwareentwickler und KI-Enthusiast. Seit vielen Jahren entwickle ich Unternehmenssoftware und beschäftige mich inzwischen fast täglich mit lokalen LLMs, KI-Agenten und leistungsfähiger KI-Hardware.

Mit LLM-Benchmark.de möchte ich eine Plattform schaffen, auf der Modelle, GPUs und Agenten objektiv und reproduzierbar miteinander verglichen werden.