Created bymario-alka.dePowered bynoob2claw.detricoma.de
Contributed byMario AlkaGoogle

Gemma-4-26B-A4B-it

Performance benchmark · measured on 27.08.2026 01:33

Benchmark-IDrun-20260826-235958-4f6662
Timebench 3 - Kombi (Prefill + Generation)MoE26BRuntime: vLLMQuantisierung: FP8
Generation258,20tok/s
Prefill16.061,90tok/s
Time to First Token813,00ms
Total duration41,34s
Concurrency5parallel
Ranking in the field
125of 344 systems

Performance benchmark · Primary metric: Generation-Speed (tok/s) · 5× concurrent

This run is better than 64 % of all comparable systems.
Generation 258,2 tok/s
+5 % vs Ø 245,2
Prefill 16.061,9 tok/s
+76 % vs Ø 9.132,6
Time to First Token 813 ms
-79 % vs Ø 3.961
Distribution in the field19 – 909 tok/s
Ø 245 Median Dieser Lauf

Wie schlägt sich dieser Benchmark mit anderen Modellen?

How does this benchmark compare on other GPUs?

Same model on different hardware · 5× concurrent · Generation (tok/s)

Hardware

GPU: NVIDIA RTX A6000 · 48 GB VRAM
CPU: AMD Ryzen Threadripper PRO 7955WX 16-Cores
RAM: 184 GB
Mainboard: ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE

Setup

Runtime: vLLM
Quantization: FP8
Model: Gemma-4-26B-A4B-it

Configuration

benchmark-konfiguration — run-20260826-235958-4f6662
# LLM-Benchmark Konfiguration # Modell : Gemma-4-26B-A4B-it # Engine : vLLM # Run-ID : run-20260826-235958-4f6662 # GPU : NVIDIA RTX A6000 # CPU : AMD Ryzen Threadripper PRO 7955WX 16-Cores # RAM : 184 GB bench@llm-benchmark:~$ vllm serve \ --model Gemma-4-26B-A4B-it
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.vllm
Modellalias?Der Name, unter dem das Modell ueber die API angesprochen wird. Genau dieser Wert muss im Request-Feld 'model' stehen.Gemma-4-26B-A4B-it
execution_typelocal

All benchmarks of this model To leaderboard

Anzeige
Model comparison

Gemma-4-26B-A4B-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

2.0761.5571.0385190,004.7339.46614.199Prefill (tok/s)Generation (tok/s)NVIDIA RTX PRO 6000 Blackwell Workstation Edition - 1.604,6 tok/s Generation, 6.276 tok/s Prefill, TTFT 4.405 ms (5 Laufe)NVIDIA RTX PRO 6000 B...NVIDIA GeForce RTX 5090 - 1.540,5 tok/s Generation, 4.006 tok/s Prefill, TTFT 4.749 ms (3 Laufe)NVIDIA GeForce RTX 50...NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition - 1.479,9 tok/s Generation, 7.172 tok/s Prefill, TTFT 3.769 ms (9 Laufe)NVIDIA RTX PRO 6000 B...NVIDIA GeForce RTX 5070 Ti - 1.009,0 tok/s Generation, 3.578 tok/s Prefill, TTFT 5.953 ms (3 Laufe)NVIDIA GeForce RTX 50...NVIDIA GeForce RTX 3090 Ti - 926,2 tok/s Generation, 3.438 tok/s Prefill, TTFT 4.863 ms (6 Laufe)NVIDIA GeForce RTX 30...AMD Radeon RX 7900 XTX - 315,5 tok/s Generation, 2.246 tok/s Prefill, TTFT 4.802 ms (6 Laufe)AMD Radeon RX 7900 XTXAMD Radeon PRO W7800 48GB - 247,3 tok/s Generation, 2.619 tok/s Prefill, TTFT 3.357 ms (8 Laufe)AMD Radeon PRO W7800 ...AMD Radeon 8060S Graphics - 79,5 tok/s Generation, 533 tok/s Prefill, TTFT 9.976 ms (9 Laufe)AMD Radeon 8060S Grap...CPU-only - 47,4 tok/s Generation, 1.385 tok/s Prefill, TTFT 1.479 ms (2 Laufe)CPU-onlyNVIDIA RTX A6000 - 841,8 tok/s Generation, 11.528 tok/s Prefill, TTFT 1.992 ms (36 Laufe) | DIESER LAUF★ NVIDIA RTX A6000
NVIDIA RTX PRO 6000 Blackwell Workstation Edition 1.604,6 tok/sNVIDIA GeForce RTX 5090 1.540,5 tok/sNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 1.479,9 tok/sNVIDIA GeForce RTX 5070 Ti 1.009,0 tok/sNVIDIA GeForce RTX 3090 Ti 926,2 tok/s★ NVIDIA RTX A6000 841,8 tok/s this runAMD Radeon RX 7900 XTX 315,5 tok/sAMD Radeon PRO W7800 48GB 247,3 tok/sAMD Radeon 8060S Graphics 79,5 tok/sCPU-only 47,4 tok/s

CPUby processor

2.0781.5581.0395190,004.7339.46614.199Prefill (tok/s)Generation (tok/s)AMD Ryzen 5 5600X 6-Core Processor - 1.604,6 tok/s Generation, 4.540 tok/s Prefill, TTFT 4.115 ms (5 Laufe)AMD Ryzen 5 5600X 6-C...AMD Ryzen 9 9950X 16-Core Processor - 1.589,1 tok/s Generation, 6.148 tok/s Prefill, TTFT 3.731 ms (3 Laufe)AMD Ryzen 9 9950X 16-...AMD Ryzen 7 5800X3D 8-Core Processor - 1.540,5 tok/s Generation, 4.006 tok/s Prefill, TTFT 4.749 ms (3 Laufe)AMD Ryzen 7 5800X3D 8...AMD Ryzen Threadripper PRO 9965WX 24-Cores - 1.479,9 tok/s Generation, 7.172 tok/s Prefill, TTFT 3.769 ms (9 Laufe)AMD Ryzen Threadrippe...AMD Ryzen Threadripper PRO 5975WX 32-Cores - 1.009,0 tok/s Generation, 2.880 tok/s Prefill, TTFT 4.065 ms (11 Laufe)AMD Ryzen Threadrippe...AMD Ryzen 9 8945HX with Radeon Graphics - 926,2 tok/s Generation, 3.620 tok/s Prefill, TTFT 6.478 ms (3 Laufe)AMD Ryzen 9 8945HX wi...AMD Ryzen Threadripper PRO 3955WX 16-Cores - 315,5 tok/s Generation, 2.594 tok/s Prefill, TTFT 4.098 ms (3 Laufe)AMD Ryzen Threadrippe...Intel(R) Xeon(R) CPU E5-4657L v2 @ 2.40GHz - 185,7 tok/s Generation, 1.898 tok/s Prefill, TTFT 5.506 ms (3 Laufe)Intel(R) Xeon(R) CPU ...AMD RYZEN AI MAX+ 395 w/ Radeon 8060S - 79,5 tok/s Generation, 533 tok/s Prefill, TTFT 9.976 ms (9 Laufe)AMD RYZEN AI MAX+ 395...Apple M6 (unified CPU/GPU SoC, ARM64, AdvSIMD) - 47,4 tok/s Generation, 1.491 tok/s Prefill, TTFT 1.402 ms (1 Lauf)Apple M6 (unified CPU...Apple M6 (unified CPU/GPU SoC) - 41,6 tok/s Generation, 1.278 tok/s Prefill, TTFT 1.556 ms (1 Lauf)Apple M6 (unified CPU...AMD Ryzen Threadripper PRO 7955WX 16-Cores - 841,8 tok/s Generation, 11.528 tok/s Prefill, TTFT 1.992 ms (36 Laufe) | DIESER LAUF★ AMD Ryzen Threadrippe...
AMD Ryzen 5 5600X 6-Core Processor 1.604,6 tok/sAMD Ryzen 9 9950X 16-Core Processor 1.589,1 tok/sAMD Ryzen 7 5800X3D 8-Core Processor 1.540,5 tok/sAMD Ryzen Threadripper PRO 9965WX 24-Cores 1.479,9 tok/sAMD Ryzen Threadripper PRO 5975WX 32-Cores 1.009,0 tok/sAMD Ryzen 9 8945HX with Radeon Graphics 926,2 tok/s★ AMD Ryzen Threadripper PRO 7955WX 16-Cores 841,8 tok/s this runAMD Ryzen Threadripper PRO 3955WX 16-Cores 315,5 tok/sIntel(R) Xeon(R) CPU E5-4657L v2 @ 2.40GHz 185,7 tok/sAMD RYZEN AI MAX+ 395 w/ Radeon 8060S 79,5 tok/sApple M6 (unified CPU/GPU SoC, ARM64, AdvSIMD) 47,4 tok/sApple M6 (unified CPU/GPU SoC) 41,6 tok/s

MBby mainboard

2.0761.5571.0385190,004.3738.74613.118Prefill (tok/s)Generation (tok/s)ASUSTeK COMPUTER INC. ROG STRIX B550-A GAMING - 1.604,6 tok/s Generation, 4.991 tok/s Prefill, TTFT 5.016 ms (5 Laufe)ASUSTeK COMPUTER INC....ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI - 1.589,1 tok/s Generation, 6.148 tok/s Prefill, TTFT 3.731 ms (3 Laufe)ASUSTeK COMPUTER INC....ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI - 1.009,0 tok/s Generation, 2.819 tok/s Prefill, TTFT 4.072 ms (14 Laufe)ASUSTeK COMPUTER INC....Meigao Innovation Technology (Shen Zhen) Co., Ltd DRFXL (MotherBoard Series) - 926,2 tok/s Generation, 3.620 tok/s Prefill, TTFT 6.478 ms (3 Laufe)Meigao Innovation Tec...ASUSTeK COMPUTER INC. PRIME A520M-K - 397,2 tok/s Generation, 3.255 tok/s Prefill, TTFT 3.248 ms (3 Laufe)ASUSTeK COMPUTER INC....Dell Inc. PowerEdge R820 - 185,7 tok/s Generation, 1.898 tok/s Prefill, TTFT 5.506 ms (3 Laufe)Dell Inc. PowerEdge R...Meigao Innovation Technology (Shen Zhen) Co., Ltd SHWSA (MS-S1 MAX) - 79,5 tok/s Generation, 533 tok/s Prefill, TTFT 9.976 ms (9 Laufe)Meigao Innovation Tec...Apple Mac mini - 47,4 tok/s Generation, 1.385 tok/s Prefill, TTFT 1.479 ms (2 Laufe)Apple Mac miniASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE - 1.479,9 tok/s Generation, 10.657 tok/s Prefill, TTFT 2.347 ms (45 Laufe) | DIESER LAUF★ ASUSTeK COMPUTER INC....
ASUSTeK COMPUTER INC. ROG STRIX B550-A GAMING 1.604,6 tok/sASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI 1.589,1 tok/s★ ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE 1.479,9 tok/s this runASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 1.009,0 tok/sMeigao Innovation Technology (Shen Zhen) Co., Ltd DRFXL (MotherBoard Series) 926,2 tok/sASUSTeK COMPUTER INC. PRIME A520M-K 397,2 tok/sDell Inc. PowerEdge R820 185,7 tok/sMeigao Innovation Technology (Shen Zhen) Co., Ltd SHWSA (MS-S1 MAX) 79,5 tok/sApple Mac mini 47,4 tok/s

ENGby engine

2.0781.5581.0395190,007.40714.81322.220Prefill (tok/s)Generation (tok/s)llama.cpp - 1.604,6 tok/s Generation, 4.092 tok/s Prefill, TTFT 4.945 ms (63 Laufe)llama.cppunbekannt - 397,2 tok/s Generation, 2.459 tok/s Prefill, TTFT 2.588 ms (4 Laufe)unbekanntlmstudio-mlx - 47,4 tok/s Generation, 1.491 tok/s Prefill, TTFT 1.402 ms (1 Lauf)lmstudio-mlxOllama - 41,6 tok/s Generation, 1.278 tok/s Prefill, TTFT 1.556 ms (1 Lauf)OllamavLLM - 841,8 tok/s Generation, 18.105 tok/s Prefill, TTFT 695 ms (18 Laufe) | DIESER LAUF★ vLLM
llama.cpp 1.604,6 tok/s★ vLLM 841,8 tok/s this rununbekannt 397,2 tok/slmstudio-mlx 47,4 tok/sOllama 41,6 tok/s

DRVby driver

2.0771.5581.0395190,002.8615.7238.584Prefill (tok/s)Generation (tok/s)unbekannt - 1.604,6 tok/s Generation, 6.933 tok/s Prefill, TTFT 3.915 ms (86 Laufe)unbekanntROCm 7.2.0 - 42,2 tok/s Generation, 71 tok/s Prefill, TTFT 609 ms (1 Lauf)ROCm 7.2.0
unbekannt 1.604,6 tok/sROCm 7.2.0 42,2 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 URLReset
⚡ Electricity price EUR/kWh
⚙️ System utilization 100 %
🖥️ Acquisition EUR
🔌 Idle 80 W
⚡ TDP 371 W
☁️ External LLM (API)
Electricity0.30 EUR/kWh
Avg power (incl. idle)371 W estimated (TDP)GPU 300 + CPU 46 + Board 25 W full load
Avg cost / hourEUR 0.11
Electricity / 1M tokensEUR 0.12
Token / kWh2.51M
Acquisition (system)EUR 8,424 full priceGPU EUR 3,000 · CPU EUR 1,399 · Board EUR 1,299 · RAM EUR 2,576 · PSU EUR 150
Electricity (2 years)–
TCO (2 years)EUR 10,374
Output tokens (2 years)16.29B
☁️ 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 (80 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-4-26B-A4B-itNVIDIA RTX A6000Gemma-4-26B-A4B-itNVIDIA RTX PRO 6000 Blackwell Workstation EditionGemma-4-26B-A4B-itNVIDIA RTX PRO 6000 Blackwell Workstation EditionGemma-4-26B-A4B-itNVIDIA GeForce RTX 5090
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.