Created bymario-alka.dePowered bygodcore.denoob2claw.detricoma.de
Contributed byMario AlkaNVIDIA

OpenReasoning-Nemotron-32B

Performance benchmark · measured on 29.07.2026 00:46

Benchmark-IDrun-20260729-032126-5c565f
Timebench 3 - Kombi (Prefill + Generation)Dense32BRuntime: llama.cppQuantisierung: Q4_K_M
Generation68,35tok/s
Prefill3.426,62tok/s
Time to First Token713,50ms
Total duration31,39s
Concurrency1parallel
Ranking in the field
163of 330 systems

Performance benchmark · Primary metric: Generation-Speed (tok/s) · 1× concurrent

This run is better than 51 % of all comparable systems.
Generation 68,4 tok/s
-40 % vs Ø 113,2
Prefill 3.426,6 tok/s
+13 % vs Ø 3.036,9
Time to First Token 714 ms
-93 % vs Ø 10.277
Distribution in the field0 – 405 tok/s
Ø 113 Median Dieser Lauf

Wie schlägt sich dieser Benchmark mit anderen Modellen?

gpt-oss-20bNVIDIA GeForce RTX 5090 · run-20260724-004608-16af9b
404,6 tok/s
NVIDIA-Nemotron-3-Nano-4BNVIDIA GeForce RTX 5090 · run-20260728-194135-d456e7
393,5 tok/s
NVIDIA-Nemotron-3-Nano-4BNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260727-181607-bffb11
393,5 tok/s
gpt-oss-20bNVIDIA GeForce RTX 5090 · run-20260724-004607-29bd16
388,9 tok/s
gpt-oss-20bNVIDIA GeForce RTX 5090 · run-20260729-032121-1c779f
388,2 tok/s
gemma-4-E2B-itNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260727-181606-6ea089
378,6 tok/s
gemma-4-E2B-itNVIDIA GeForce RTX 5090 · run-20260728-184455-e8b129
377,6 tok/s
NVIDIA-Nemotron-3-Nano-4B3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260728-194136-ac388e
362,3 tok/s
NVIDIA-Nemotron-3-Nano-4B3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260728-194136-5e5085
361,4 tok/s
Nemotron-3-Nano-Omni-30B-A3B-ReasoningNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032119-4852bf
357,1 tok/s
NVIDIA-Nemotron-3-Nano-30B-A3BNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032121-643b00
357,0 tok/s
Nemotron-Cascade-2-30B-A3BNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032120-4e3476
356,8 tok/s
NVIDIA-Nemotron-3-Nano-4B3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260727-181606-a2f36d
356,7 tok/s
NVIDIA-Nemotron-3-Nano-4B3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260728-194136-1be57b
355,6 tok/s
OpenReasoning-Nemotron-32B this runNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032126-5c565f
68,4 tok/s

How does this benchmark compare on other GPUs?

Same model on different hardware · 1× concurrent · Generation (tok/s)

Hardware

GPU: NVIDIA RTX PRO 6000 Blackwell Workstation Edition · 96 GB VRAM
CPU: AMD Ryzen 9 9950X 16-Core Processor
RAM: 92 GB
Mainboard: ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI

Setup

Runtime: llama.cpp
Quantization: Q4_K_M
Model: OpenReasoning-Nemotron-32B

Configuration

benchmark-konfiguration — run-20260729-032126-5c565f
# LLM-Benchmark Konfiguration # Modell : OpenReasoning-Nemotron-32B # Engine : llama.cpp # Run-ID : run-20260729-032126-5c565f # GPU : NVIDIA RTX PRO 6000 Blackwell Workstation Edition # CPU : AMD Ryzen 9 9950X 16-Core Processor # RAM : 92 GB bench@llm-benchmark:~$ /home/godcore/llama.cpp/build/bin/llama-server \ -m /home/godcore/.cache/huggingface/hub/models--gabriellarson--OpenReasoning-Nemotron-32B-GGUF/snapshots/56ebd68caf9cf4b656f46e5aad225420dd6b71d4/OpenReasoning-Nemotron-32B-Q4_K_M.gguf \ --alias OpenReasoning-Nemotron-32B \ --host 0.0.0.0 \ --port 8000 \ -ngl 999 \ -c 16384 \ -np 4 \ --jinja
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.OpenReasoning-Nemotron-32B
Kontextlaenge?Maximale Anzahl Tokens (Eingabe + erzeugte Ausgabe zusammen), die das Modell pro Anfrage verarbeiten kann.16384
Modellpfad?Pfad zur GGUF-Modelldatei, die geladen und ausgeliefert wird./home/godcore/.cache/huggingface/hub/models--gabriellarson--OpenReasoning-Nemotron-32B-GGUF/snapshots/56ebd68caf9cf4b656f46e5aad225420dd6b71d4/OpenReasoning-Nemotron-32B-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
Kontext?Groesse des Kontextfensters in Token. 0 = der beim Training verwendete Kontext des Modells.16384
np4

All benchmarks of this model To leaderboard

Anzeige
Model comparison

OpenReasoning-Nemotron-32B 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

6244683121560,001.8603.7215.581Prefill (tok/s)Generation (tok/s)AMD Radeon AI PRO R9700 - 64,3 tok/s Generation, 3.816 tok/s Prefill, TTFT 5.428 ms (2 Laufe)AMD Radeon AI PRO R97...NVIDIA GeForce RTX 5070 Ti - 49,9 tok/s Generation, 621 tok/s Prefill, TTFT 94.653 ms (3 Laufe)NVIDIA GeForce RTX 50...AMD Radeon 8060S Graphics - 21,1 tok/s Generation, 429 tok/s Prefill, TTFT 15.251 ms (2 Laufe)AMD Radeon 8060S Grap...CPU-only - 1,1 tok/s Generation, 8 tok/s Prefill, TTFT 266.668 ms (1 Lauf)CPU-onlyNVIDIA RTX PRO 6000 Blackwell Workstation Edition - 480,0 tok/s Generation, 4.502 tok/s Prefill, TTFT 9.693 ms (3 Laufe) | DIESER LAUF★ NVIDIA RTX PRO 6000 B...
★ NVIDIA RTX PRO 6000 Blackwell Workstation Edition 480,0 tok/s this runAMD Radeon AI PRO R9700 64,3 tok/sNVIDIA GeForce RTX 5070 Ti 49,9 tok/sAMD Radeon 8060S Graphics 21,1 tok/sCPU-only 1,1 tok/s

CPUby processor

6244683121560,001.8603.7215.581Prefill (tok/s)Generation (tok/s)AMD Ryzen Threadripper PRO 7955WX 16-Cores - 64,3 tok/s Generation, 3.816 tok/s Prefill, TTFT 5.428 ms (2 Laufe)AMD Ryzen Threadrippe...AMD Ryzen Threadripper PRO 5975WX 32-Cores - 49,9 tok/s Generation, 621 tok/s Prefill, TTFT 94.653 ms (3 Laufe)AMD Ryzen Threadrippe...AMD RYZEN AI MAX+ 395 w/ Radeon 8060S - 21,1 tok/s Generation, 429 tok/s Prefill, TTFT 15.251 ms (2 Laufe)AMD RYZEN AI MAX+ 395...Intel(R) Xeon(R) CPU E5-4620 v2 @ 2.60GHz - 1,1 tok/s Generation, 8 tok/s Prefill, TTFT 266.668 ms (1 Lauf)Intel(R) Xeon(R) CPU ...AMD Ryzen 9 9950X 16-Core Processor - 480,0 tok/s Generation, 4.502 tok/s Prefill, TTFT 9.693 ms (3 Laufe) | DIESER LAUF★ AMD Ryzen 9 9950X 16-...
★ AMD Ryzen 9 9950X 16-Core Processor 480,0 tok/s this runAMD Ryzen Threadripper PRO 7955WX 16-Cores 64,3 tok/sAMD Ryzen Threadripper PRO 5975WX 32-Cores 49,9 tok/sAMD RYZEN AI MAX+ 395 w/ Radeon 8060S 21,1 tok/sIntel(R) Xeon(R) CPU E5-4620 v2 @ 2.60GHz 1,1 tok/s

MBby mainboard

6244683121560,001.8603.7215.581Prefill (tok/s)Generation (tok/s)ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE - 64,3 tok/s Generation, 3.816 tok/s Prefill, TTFT 5.428 ms (2 Laufe)ASUSTeK COMPUTER INC....ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI - 49,9 tok/s Generation, 621 tok/s Prefill, TTFT 94.653 ms (3 Laufe)ASUSTeK COMPUTER INC....Bosgame AXB35-02 (BeyondMax Series) - 21,1 tok/s Generation, 429 tok/s Prefill, TTFT 15.251 ms (2 Laufe)Bosgame AXB35-02 (Bey...Dell Inc. PowerEdge R820 - 1,1 tok/s Generation, 8 tok/s Prefill, TTFT 266.668 ms (1 Lauf)Dell Inc. PowerEdge R...ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI - 480,0 tok/s Generation, 4.502 tok/s Prefill, TTFT 9.693 ms (3 Laufe) | DIESER LAUF★ ASUSTeK COMPUTER INC....
★ ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI 480,0 tok/s this runASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE 64,3 tok/sASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 49,9 tok/sBosgame AXB35-02 (BeyondMax Series) 21,1 tok/sDell Inc. PowerEdge R820 1,1 tok/s

ENGby engine

5285044804564322.0392.1262.2132.300Prefill (tok/s)Generation (tok/s)llama.cpp - 480,0 tok/s Generation, 2.169 tok/s Prefill, TTFT 56.460 ms (11 Laufe) | DIESER LAUF★ llama.cpp
★ llama.cpp 480,0 tok/s this run

DRVby driver

5285044804564322.0392.1262.2132.300Prefill (tok/s)Generation (tok/s)unbekannt - 480,0 tok/s Generation, 2.169 tok/s Prefill, TTFT 56.460 ms (11 Laufe)unbekannt
unbekannt 480,0 tok/s
💰 Economics

Economics of this run

Operating cost, TCO and comparison with the next-best runs of the same model at identical concurrency (1× concurrent). Methodology →

⚙️ ConfigurationAll metrics and charts below follow these settings – based on a 24-month runtime.Save to URLReset
⚡ Electricity price EUR/kWh
⚙️ System utilization 100 %
🖥️ Acquisition EUR
🔌 Idle 70 W
⚡ TDP 644 W
☁️ External LLM (API)
Electricity0.30 EUR/kWh
Avg power (incl. idle)644 W estimated (TDP)GPU 600 + CPU 29 + Board 15 W full load
Avg cost / hourEUR 0.19
Electricity / 1M tokensEUR 0.78
Token / kWh382.23K
Acquisition (system)EUR 15,248 full priceGPU EUR 13,000 · CPU EUR 649 · Board EUR 499 · RAM EUR 920 · PSU EUR 180
Electricity (2 years)
TCO (2 years)EUR 18,632
Output tokens (2 years)4.31B
☁️ 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 (70 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

OpenReasoning-Nemotron-32BNVIDIA RTX PRO 6000 Blackwell Workstation EditionOpenReasoning-Nemotron-32BAMD Radeon 8060S GraphicsOpenReasoning-Nemotron-32BNVIDIA GeForce RTX 5070 TiOpenReasoning-Nemotron-32BKeine GPU (CPU-only)
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)

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.