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

Magistral-Small-2509

Performance benchmark · measured on 18.08.2026 08:46

Benchmark-IDrun-20260818-090950-dbac68
Timebench 3 - Kombi (Prefill + Generation)Dense24BRuntime: llama.cppQuantisierung: Q8_0
Generation92,32tok/s
Prefill1.519,76tok/s
Time to First Token8.028,50ms
Total duration129,54s
Concurrency5parallel
Ranking in the field
485of 841 systems

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

This run is better than 42 % of all comparable systems.
Generation 92,3 tok/s
-66 % vs Ø 274,3
Prefill 1.519,8 tok/s
-61 % vs Ø 3.891,1
Time to First Token 8.029 ms
-81 % vs Ø 41.842
Distribution in the field0 – 1.349 tok/s
Ø 274 Median Dieser Lauf

Wie schlägt sich dieser Benchmark mit anderen Modellen?

gemma-4-E2B-itNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260727-181606-0adf6d
1.349,3 tok/s
gemma-4-E2B-itNVIDIA GeForce RTX 5090 · run-20260728-184455-5b4937
1.301,9 tok/s
gemma-4-E2B-it3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260728-184456-4111a9
1.195,7 tok/s
Nemotron-3-Nano-4BNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260727-181607-30d864
1.188,7 tok/s
Nemotron-3-Nano-4BNVIDIA GeForce RTX 5090 · run-20260728-194135-5432cd
1.164,8 tok/s
gpt-oss-20bNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260728-140954-9e81f9
1.136,3 tok/s
gpt-oss-20bNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260730-035052-09627f
1.135,0 tok/s
gpt-oss-20bNVIDIA GeForce RTX 5090 · run-20260729-032121-058a31
1.113,5 tok/s
Nemotron-3-Nano-4B3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260728-194136-836a91
1.090,6 tok/s
Nemotron-3-Nano-4B3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260727-181606-8b381e
1.077,4 tok/s
Nemotron-3-Nano-4B3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260728-194136-b8a818
1.071,0 tok/s
Nemotron-3-Nano-4B3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260728-194136-038e92
1.051,1 tok/s
gpt-oss-20b3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260729-032121-ffb18a
1.015,5 tok/s
Nemotron-3-Nano-30B-A3BNVIDIA RTX PRO 6000 Blackwell Workstation Edition · run-20260729-032121-305ad6
1.007,8 tok/s
Magistral-Small-2509 this run2× AMD Radeon PRO W7900 Dual Slot · run-20260818-090950-dbac68
92,3 tok/s

How does this benchmark compare on other GPUs?

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

Hardware

GPU: 2x AMD Radeon PRO W7900 Dual Slot · 48 GB VRAM
CPU: AMD Ryzen Threadripper PRO 5975WX 32-Cores
RAM: 247 GB
Mainboard: ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI

Setup

Runtime: llama.cpp
Quantization: Q8_0
Model: Magistral-Small-2509

Configuration

benchmark-konfiguration — run-20260818-090950-dbac68
# LLM-Benchmark Konfiguration # Modell : Magistral-Small-2509 # Engine : llama.cpp # Run-ID : run-20260818-090950-dbac68 # GPU : 2x AMD Radeon PRO W7900 Dual Slot # CPU : AMD Ryzen Threadripper PRO 5975WX 32-Cores # RAM : 247 GB bench@llm-benchmark:~$ /home/godcore/llama.cpp/build_hip/bin/llama-server \ -m /home/godcore/bench_scratch/mistralai_Magistral-Small-2509-Q8_0.gguf \ --alias Magistral-Small-2509 \ -ngl 999 \ -fa on \ -sm layer \ --tensor-split 1,1 \ -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.Magistral-Small-2509
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/bench_scratch/mistralai_Magistral-Small-2509-Q8_0.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
faon
Split-Mode?Verteilung ueber mehrere GPUs: none (nur eine GPU), layer (Layer aufteilen) oder row (Tensoren zeilenweise).layer
Tensor-Split?Verhaeltnis, in dem die Modell-Layer auf mehrere GPUs verteilt werden, z.B. 3,1.1,1
Kontext?Groesse des Kontextfensters in Token. 0 = der beim Training verwendete Kontext des Modells.49152
np12

All benchmarks of this model To leaderboard

Anzeige
Model comparison

Magistral-Small-2509 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

9537154772380,001.6063.2124.817Prefill (tok/s)Generation (tok/s)NVIDIA GeForce RTX 5090 - 737,0 tok/s Generation, 128 tok/s Prefill, TTFT 3.689 ms (3 Laufe)NVIDIA GeForce RTX 50...NVIDIA RTX PRO 6000 Blackwell Workstation Edition - 726,9 tok/s Generation, 3.891 tok/s Prefill, TTFT 2.194 ms (6 Laufe)NVIDIA RTX PRO 6000 B...NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition - 637,2 tok/s Generation, 149 tok/s Prefill, TTFT 3.798 ms (12 Laufe)NVIDIA RTX PRO 6000 B...NVIDIA GeForce RTX 3090 Ti - 353,8 tok/s Generation, 1.856 tok/s Prefill, TTFT 4.508 ms (6 Laufe)NVIDIA GeForce RTX 30...AMD Radeon AI PRO R9700 - 230,0 tok/s Generation, 910 tok/s Prefill, TTFT 6.435 ms (15 Laufe)AMD Radeon AI PRO R97...NVIDIA GeForce RTX 5070 Ti - 159,0 tok/s Generation, 78 tok/s Prefill, TTFT 11.931 ms (3 Laufe)NVIDIA GeForce RTX 50...Intel Arc Pro B70 - 37,9 tok/s Generation, 43 tok/s Prefill, TTFT 65.064 ms (3 Laufe)Intel Arc Pro B70NVIDIA Tesla P100 PCIe 16GB - 24,7 tok/s Generation, 189 tok/s Prefill, TTFT 35.758 ms (2 Laufe)NVIDIA Tesla P100 PCI...AMD Radeon PRO W7900 Dual Slot - 141,7 tok/s Generation, 1.590 tok/s Prefill, TTFT 8.535 ms (12 Laufe) | DIESER LAUF★ AMD Radeon PRO W7900 ...
NVIDIA GeForce RTX 5090 737,0 tok/sNVIDIA RTX PRO 6000 Blackwell Workstation Edition 726,9 tok/sNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 637,2 tok/sNVIDIA GeForce RTX 3090 Ti 353,8 tok/sAMD Radeon AI PRO R9700 230,0 tok/sNVIDIA GeForce RTX 5070 Ti 159,0 tok/s★ AMD Radeon PRO W7900 Dual Slot 141,7 tok/s this runIntel Arc Pro B70 37,9 tok/sNVIDIA Tesla P100 PCIe 16GB 24,7 tok/s

CPUby processor

9507134752380,001.6023.2054.807Prefill (tok/s)Generation (tok/s)AMD Ryzen 7 5800X3D 8-Core Processor - 737,0 tok/s Generation, 128 tok/s Prefill, TTFT 3.689 ms (3 Laufe)AMD Ryzen 7 5800X3D 8...AMD Ryzen 9 9950X 16-Core Processor - 726,9 tok/s Generation, 3.891 tok/s Prefill, TTFT 2.194 ms (6 Laufe)AMD Ryzen 9 9950X 16-...AMD Ryzen Threadripper PRO 9965WX 24-Cores - 637,2 tok/s Generation, 149 tok/s Prefill, TTFT 3.798 ms (12 Laufe)AMD Ryzen Threadrippe...AMD Ryzen 9 8945HX with Radeon Graphics - 353,8 tok/s Generation, 1.856 tok/s Prefill, TTFT 4.508 ms (6 Laufe)AMD Ryzen 9 8945HX wi...AMD Ryzen Threadripper PRO 7955WX 16-Cores - 230,0 tok/s Generation, 910 tok/s Prefill, TTFT 6.435 ms (15 Laufe)AMD Ryzen Threadrippe...AMD Ryzen 9 7945HX with Radeon Graphics - 37,9 tok/s Generation, 102 tok/s Prefill, TTFT 53.342 ms (5 Laufe)AMD Ryzen 9 7945HX wi...AMD Ryzen Threadripper PRO 5975WX 32-Cores - 159,0 tok/s Generation, 1.288 tok/s Prefill, TTFT 9.214 ms (15 Laufe) | DIESER LAUF★ AMD Ryzen Threadrippe...
AMD Ryzen 7 5800X3D 8-Core Processor 737,0 tok/sAMD Ryzen 9 9950X 16-Core Processor 726,9 tok/sAMD Ryzen Threadripper PRO 9965WX 24-Cores 637,2 tok/sAMD Ryzen 9 8945HX with Radeon Graphics 353,8 tok/sAMD Ryzen Threadripper PRO 7955WX 16-Cores 230,0 tok/s★ AMD Ryzen Threadripper PRO 5975WX 32-Cores 159,0 tok/s this runAMD Ryzen 9 7945HX with Radeon Graphics 37,9 tok/s

MBby mainboard

9537154772380,001.6063.2124.817Prefill (tok/s)Generation (tok/s)ASUSTeK COMPUTER INC. ROG STRIX B550-A GAMING - 737,0 tok/s Generation, 128 tok/s Prefill, TTFT 3.689 ms (3 Laufe)ASUSTeK COMPUTER INC....ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI - 726,9 tok/s Generation, 3.891 tok/s Prefill, TTFT 2.194 ms (6 Laufe)ASUSTeK COMPUTER INC....ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE - 637,2 tok/s Generation, 572 tok/s Prefill, TTFT 5.263 ms (27 Laufe)ASUSTeK COMPUTER INC....Meigao Innovation Technology (Shen Zhen) Co., Ltd DRFXL (MotherBoard Series) - 353,8 tok/s Generation, 1.856 tok/s Prefill, TTFT 4.508 ms (6 Laufe)Meigao Innovation Tec...Shenzhen Meigao Electronic Equipment Co.,Ltd DRFXI (MotherBoard Series) - 37,9 tok/s Generation, 43 tok/s Prefill, TTFT 65.064 ms (3 Laufe)Shenzhen Meigao Elect...Shenzhen Meigao Electronic Equipment Co.,Ltd F1FXM (DeskMini Series) - 24,7 tok/s Generation, 189 tok/s Prefill, TTFT 35.758 ms (2 Laufe)Shenzhen Meigao Elect...ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI - 159,0 tok/s Generation, 1.288 tok/s Prefill, TTFT 9.214 ms (15 Laufe) | DIESER LAUF★ ASUSTeK COMPUTER INC....
ASUSTeK COMPUTER INC. ROG STRIX B550-A GAMING 737,0 tok/sASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI 726,9 tok/sASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE 637,2 tok/sMeigao Innovation Technology (Shen Zhen) Co., Ltd DRFXL (MotherBoard Series) 353,8 tok/s★ ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 159,0 tok/s this runShenzhen Meigao Electronic Equipment Co.,Ltd DRFXI (MotherBoard Series) 37,9 tok/sShenzhen Meigao Electronic Equipment Co.,Ltd F1FXM (DeskMini Series) 24,7 tok/s

ENGby engine

9527144762380,002.1644.3276.491Prefill (tok/s)Generation (tok/s)vLLM - 726,9 tok/s Generation, 5.326 tok/s Prefill, TTFT 2.928 ms (6 Laufe)vLLMunbekannt - 28,9 tok/s Generation, 1.586 tok/s Prefill, TTFT 1.672 ms (3 Laufe)unbekanntllama.cpp - 737,0 tok/s Generation, 631 tok/s Prefill, TTFT 10.863 ms (53 Laufe) | DIESER LAUF★ llama.cpp
★ llama.cpp 737,0 tok/s this runvLLM 726,9 tok/sunbekannt 28,9 tok/s

DRVby driver

9527144762380,006531.3061.958Prefill (tok/s)Generation (tok/s)unbekannt - 737,0 tok/s Generation, 1.165 tok/s Prefill, TTFT 7.109 ms (56 Laufe)unbekanntIntel 26.18.38308.4 - 37,9 tok/s Generation, 43 tok/s Prefill, TTFT 65.064 ms (3 Laufe)Intel 26.18.38308.4AMD 7.0.0-27-generic - 28,9 tok/s Generation, 1.586 tok/s Prefill, TTFT 1.672 ms (3 Laufe)AMD 7.0.0-27-generic
unbekannt 737,0 tok/sIntel 26.18.38308.4 37,9 tok/sAMD 7.0.0-27-generic 28,9 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 10 W
⚡ TDP 600 W
☁️ External LLM (API)
Electricity0.30 EUR/kWh
Avg power (incl. idle)600 W estimated (TDP)GPU 590 + Board 10 W full load
Avg cost / hourEUR 0.18
Electricity / 1M tokensEUR 0.54
Token / kWh553.92K
Acquisition (system)EUR 2,156 missingRAM EUR 1,976 · PSU EUR 180
Electricity (2 years)
TCO (2 years)EUR 5,310
Output tokens (2 years)5.82B
☁️ 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 (10 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

Magistral-Small-25092x AMD Radeon PRO W7900 Dual SlotMagistral-Small-2509NVIDIA RTX PRO 6000 Blackwell Workstation EditionMagistral-Small-2509NVIDIA RTX PRO 6000 Blackwell Workstation EditionMagistral-Small-2509NVIDIA 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.