# LLM-Benchmark Konfiguration # Modell : Devstral-Small-2-24B-Instruct-2512 # Engine : vLLM # Run-ID : run-20260724-004603-630e7a # GPU : NVIDIA RTX PRO 6000 Blackwell Workstation Edition # CPU : AMD Ryzen 9 9950X 16-Core Processor # RAM : 92 GB bench@llm-benchmark:~$ /home/godcore/minimax-vllm-nightly/.venv/bin/python /home/godcore/minimax-vllm-nightly/.venv/bin/vllm serve cyankiwi/Devstral-Small-2-24B-Instruct-2512-AWQ-4bit \ --served-model-name Devstral-Small-2-24B-Instruct-2512 \ --dtype auto \ --max-model-len 8192 \ --gpu-memory-utilization 0.90 \ --trust-remote-code \ --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.	vllm
Modellalias?Der Name, unter dem das Modell ueber die API angesprochen wird. Genau dieser Wert muss im Request-Feld 'model' stehen.	Devstral-Small-2-24B-Instruct-2512
Kontextlaenge?Maximale Anzahl Tokens (Eingabe + erzeugte Ausgabe zusammen), die das Modell pro Anfrage verarbeiten kann.	8192
Alias?Anzeigename des Modells nach aussen (served model name), unabhaengig vom Dateinamen.	Devstral-Small-2-24B-Instruct-2512
Dtype?Zahlenformat der Modellgewichte bei der Berechnung (z.B. auto, float16, bfloat16). 'auto' waehlt automatisch das vom Modell empfohlene Format.	auto
Kontext?Groesse des Kontextfensters in Token. 0 = der beim Training verwendete Kontext des Modells.	8192
GPU-Speicher?Anteil des GPU-Speichers (0 bis 1), den vLLM belegen darf. 0.92 = 92 %. Hoeher = mehr Platz fuer den KV-Cache (mehr/laengere parallele Anfragen), aber groesseres Risiko fuer 'Out of Memory'.	0.90

All benchmarks of this model To leaderboard

Model comparison

Devstral-Small-2-24B-Instruct-2512 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 Workstation Edition 725,6 tok/s this runNVIDIA GeForce RTX 5090 620,3 tok/sNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 569,7 tok/sNVIDIA RTX A6000 536,7 tok/sNVIDIA GeForce RTX 3090 Ti 331,6 tok/sAMD Radeon AI PRO R9700 206,3 tok/sAMD Radeon PRO W7900 Dual Slot 152,1 tok/sAMD Radeon PRO W7800 48GB 138,4 tok/sNVIDIA GeForce RTX 5070 Ti 108,5 tok/sNVIDIA Tesla P100 PCIe 16GB 23,3 tok/sAMD Radeon 8060S Graphics 8,9 tok/s

CPUby processor

★ AMD Ryzen 9 9950X 16-Core Processor 725,6 tok/s this runAMD Ryzen 7 5800X3D 8-Core Processor 620,3 tok/sAMD Ryzen Threadripper PRO 9965WX 24-Cores 569,7 tok/sAMD Ryzen Threadripper PRO 7955WX 16-Cores 536,7 tok/sAMD Ryzen 9 8945HX with Radeon Graphics 331,6 tok/sAMD Ryzen 5 5600X 6-Core Processor 247,7 tok/sAMD Ryzen Threadripper PRO 5975WX 32-Cores 152,1 tok/sAMD Ryzen 9 7945HX with Radeon Graphics 23,3 tok/sAMD RYZEN AI MAX+ 395 w/ Radeon 8060S 8,9 tok/s

MBby mainboard

★ ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI 725,6 tok/s this runASUSTeK COMPUTER INC. ROG STRIX B550-A GAMING 620,3 tok/sASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE 569,7 tok/sMeigao Innovation Technology (Shen Zhen) Co., Ltd DRFXL (MotherBoard Series) 331,6 tok/sASUSTeK COMPUTER INC. PRIME A520M-K 247,7 tok/sASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 152,1 tok/sShenzhen Meigao Electronic Equipment Co.,Ltd F1FXM (DeskMini Series) 23,3 tok/sMeigao Innovation Technology (Shen Zhen) Co., Ltd SHWSA (MS-S1 MAX) 8,9 tok/sBosgame AXB35-02 (BeyondMax Series) 8,9 tok/s

ENGby engine

★ vLLM 725,6 tok/s this runllama.cpp 645,1 tok/sunbekannt 247,7 tok/s

DRVby driver

unbekannt 725,6 tok/sAMD 7.0.0-27-generic 29,4 tok/sROCm 7.2.0 8,9 tok/s

💰 Economics

Economics of this run

Operating cost, TCO and comparison with the next-best runs of the same model at identical concurrency (1× 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 70 W

⚡ TDP 644 W

☁️ External LLM (API) show

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

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 0.55

Token / kWh541.72K

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)6.11B

☁️ 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

	Devstral-Small-2-24B-Instruct-2512NVIDIA RTX PRO 6000 Blackwell Workstation Edition	Devstral-Small-2-24B-Instruct-2512NVIDIA GeForce RTX 5090	Devstral-Small-2-24B-Instruct-2512NVIDIA RTX PRO 6000 Blackwell Workstation Edition	Devstral-Small-2-24B-Instruct-25123x 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)	–	–	–	–

NVIDIA RTX PRO 6000 Blackwell Workstation Edition · this run 97 tok/s NVIDIA GeForce RTX 5090 95 tok/s NVIDIA RTX PRO 6000 Blackwell Workstation Edition 94 tok/s 3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 89 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.