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

gemma-4-31B-it

Performance benchmark · measured on 27.07.2026 18:09

Benchmark-IDrun-20260727-181607-1949e0
Timebench 3 - Kombi (Prefill + Generation)Dense31BRuntime: llama.cppQuantisierung: INT8
Generation394,47tok/s
Prefill2.419,92tok/s
Time to First Token29.759,00ms
Total duration173,62s
Concurrency10parallel
Ranking in the field
6of 16 systems

Performance benchmark · Primary metric: Generation-Speed (tok/s) · 10× concurrent

This run is better than 67 % of all comparable systems.
Generation 394,5 tok/s
+1 % vs Ø 391,7
Prefill 2.419,9 tok/s
-57 % vs Ø 5.646,0
Time to First Token 29.759 ms
+56 % vs Ø 19.059
Distribution in the field55 – 1.897 tok/s
Ø 392 Median Dieser Lauf

Wie schlägt sich dieser Benchmark mit anderen Modellen?

NVIDIA-Nemotron-3-Nano-4B3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260727-181606-e0eafd
1.897,0 tok/s
gpt-oss-20b3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260724-082940-10761f
1.238,3 tok/s
GLM-4.5-Air3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260727-193100-9a2d59
615,3 tok/s
Devstral-Small-25073× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260727-181606-59af91
572,2 tok/s
gemma-4-31B-it same model3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260727-193100-f3344f
397,5 tok/s
gemma-4-31B-it this run3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260727-181607-1949e0
394,5 tok/s
gemma-4-E2B-it3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260727-155936-b597f5
266,1 tok/s
gemma-4-E4B-it3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260727-181605-64c902
137,2 tok/s
command-a-reasoning-08-20253× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260727-181606-ca0b50
134,9 tok/s
gpt-oss-20b3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260724-082952-499c95
109,1 tok/s
gpt-oss-20b3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260725-031737-ac6fbc
90,4 tok/s
gpt-oss-20b3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260724-082951-9eed96
90,3 tok/s
gpt-oss-20b3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260725-031737-18d841
90,1 tok/s
gpt-oss-20b3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260724-082952-aeb5a6
89,8 tok/s

How does this benchmark compare on other GPUs?

Same model on different hardware · 10× concurrent · Generation (tok/s)

Hardware

GPU: 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 96 GB VRAM
CPU: AMD Ryzen Threadripper PRO 9965WX 24-Cores
RAM: 125 GB
Mainboard: ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE

Setup

Runtime: llama.cpp
Quantization: INT8
Model: gemma-4-31B-it

Configuration

benchmark-konfiguration — run-20260727-181607-1949e0
# LLM-Benchmark Konfiguration # Modell : gemma-4-31B-it # Engine : llama.cpp # Run-ID : run-20260727-181607-1949e0 # GPU : 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition # CPU : AMD Ryzen Threadripper PRO 9965WX 24-Cores # RAM : 125 GB bench@llm-benchmark:~$ /home/godcore/llama.cpp/build/bin/llama-server \ -m /home/godcore/.cache/huggingface/hub/models--ReadyArt--gemma-4-31B-it-scotoma-GGUF/snapshots/273b4b513d43680abdad00a219bc938032547f12/gemma-4-31B-scotoma-Q4_K_M.gguf \ --alias gemma-4-31b-it-int8 \ --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.gemma-4-31b-it-int8
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--ReadyArt--gemma-4-31B-it-scotoma-GGUF/snapshots/273b4b513d43680abdad00a219bc938032547f12/gemma-4-31B-scotoma-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

Model comparison

gemma-4-31B-it 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

5133852561280,009231.8462.770Prefill (tok/s)Generation (tok/s)NVIDIA GeForce RTX 5070 Ti - 42,0 tok/s Generation, 364 tok/s Prefill, TTFT 71.670 ms (6 Laufe)NVIDIA GeForce RTX 50...NVIDIA GeForce RTX 3090 Ti - 19,5 tok/s Generation, 345 tok/s Prefill, TTFT 88.782 ms (6 Laufe)NVIDIA GeForce RTX 30...NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition - 397,5 tok/s Generation, 2.284 tok/s Prefill, TTFT 13.330 ms (6 Laufe) | DIESER LAUF★ NVIDIA RTX PRO 6000 B...
★ NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 397,5 tok/s this runNVIDIA GeForce RTX 5070 Ti 42,0 tok/sNVIDIA GeForce RTX 3090 Ti 19,5 tok/s

CPUby processor

5133852561280,009231.8462.770Prefill (tok/s)Generation (tok/s)AMD Ryzen Threadripper PRO 5975WX 32-Cores - 42,0 tok/s Generation, 364 tok/s Prefill, TTFT 71.670 ms (6 Laufe)AMD Ryzen Threadrippe...AMD Ryzen 9 8945HX with Radeon Graphics - 19,5 tok/s Generation, 345 tok/s Prefill, TTFT 88.782 ms (6 Laufe)AMD Ryzen 9 8945HX wi...AMD Ryzen Threadripper PRO 9965WX 24-Cores - 397,5 tok/s Generation, 2.284 tok/s Prefill, TTFT 13.330 ms (6 Laufe) | DIESER LAUF★ AMD Ryzen Threadrippe...
★ AMD Ryzen Threadripper PRO 9965WX 24-Cores 397,5 tok/s this runAMD Ryzen Threadripper PRO 5975WX 32-Cores 42,0 tok/sAMD Ryzen 9 8945HX with Radeon Graphics 19,5 tok/s

MBby mainboard

5133852561280,009231.8462.770Prefill (tok/s)Generation (tok/s)ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI - 42,0 tok/s Generation, 364 tok/s Prefill, TTFT 71.670 ms (6 Laufe)ASUSTeK COMPUTER INC....Meigao Innovation Technology (Shen Zhen) Co., Ltd DRFXL (MotherBoard Series) - 19,5 tok/s Generation, 345 tok/s Prefill, TTFT 88.782 ms (6 Laufe)Meigao Innovation Tec...ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE - 397,5 tok/s Generation, 2.284 tok/s Prefill, TTFT 13.330 ms (6 Laufe) | DIESER LAUF★ ASUSTeK COMPUTER INC....
★ ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE 397,5 tok/s this runASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 42,0 tok/sMeigao Innovation Technology (Shen Zhen) Co., Ltd DRFXL (MotherBoard Series) 19,5 tok/s

ENGby engine

4374173973783589389781.0181.058Prefill (tok/s)Generation (tok/s)llama.cpp - 397,5 tok/s Generation, 998 tok/s Prefill, TTFT 57.927 ms (18 Laufe) | DIESER LAUF★ llama.cpp
★ llama.cpp 397,5 tok/s this run

DRVby driver

4374173973783589389781.0181.058Prefill (tok/s)Generation (tok/s)unbekannt - 397,5 tok/s Generation, 998 tok/s Prefill, TTFT 57.927 ms (18 Laufe)unbekannt
unbekannt 397,5 tok/s
💰 Economics

Economics of this run

Operating cost, TCO and comparison with the next-best runs of the same model at identical concurrency (10× 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 165 W
⚡ TDP 983 W
☁️ External LLM (API)
Electricity0.30 EUR/kWh
Avg power (incl. idle)983 W estimated (TDP)GPU 900 + CPU 57 + Board 25 W full load
Avg cost / hourEUR 0.29
Electricity / 1M tokensEUR 0.21
Token / kWh1.45M
Acquisition (system)EUR 45,748 full priceGPU EUR 39,000 · CPU EUR 3,499 · Board EUR 1,299 · RAM EUR 1,750 · PSU EUR 200
Electricity (2 years)
TCO (2 years)EUR 50,912
Output tokens (2 years)24.88B
☁️ 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 (165 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

gemma-4-31B-it3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Editiongemma-4-31B-it3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Editiongemma-4-31B-itNVIDIA GeForce RTX 5070 Tigemma-4-31B-itNVIDIA GeForce RTX 5070 Ti
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.