Created bymario-alka.dePowered bynoob2claw.detricoma.de
Contributed byMario AlkaQwen (Alibaba)

Qwen2.5-3B-Instruct

Performance benchmark · measured on 22.09.2026 09:00

Benchmark-IDrun-20260922-071040-5f4866
Timebench 3 - Kombi (Prefill + Generation)Dense3BRuntime: llama.cppQuantisierung: Q8_0
Generation289,42tok/s
Prefill10.849,24tok/s
Time to First Token1.211,50ms
Total duration38,47s
Concurrency5parallel
Ranking in the field
6of 244 systems

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

This run is better than 98 % of all comparable systems.
Generation 289,4 tok/s
+209 % vs Ø 93,6
Prefill 10.849,2 tok/s
+388 % vs Ø 2.223,4
Time to First Token 1.212 ms
-98 % vs Ø 68.160
Distribution in the field0 – 449 tok/s
Ø 94 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: AMD Radeon AI PRO R9700 · 32 GB VRAM
CPU: AMD Ryzen 3 3100 4-Core Processor
RAM: 31 GB
Mainboard: ASUSTeK COMPUTER INC. ROG STRIX X570-F GAMING

Setup

Runtime: llama.cpp
Quantization: Q8_0
Model: Qwen2.5-3B-Instruct

Configuration

benchmark-konfiguration — run-20260922-071040-5f4866
# LLM-Benchmark Konfiguration # Modell : Qwen2.5-3B-Instruct # Engine : llama.cpp # Run-ID : run-20260922-071040-5f4866 # GPU : AMD Radeon AI PRO R9700 # CPU : AMD Ryzen 3 3100 4-Core Processor # RAM : 31 GB bench@llm-benchmark:~$ llama-server \ -m /home/godcore/models/qwen2.5-3b-instruct-q8_0.gguf \ --alias Qwen2.5-3B-Instruct \ -ngl 999 \ -fa on \ -c 49152 \ -np 12 '(AMD' Radeon AI PRO R9700, ROCm 7.2.4, 'HIP_VISIBLE_DEVICES=0)'
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.Qwen2.5-3B-Instruct
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/models/qwen2.5-3b-instruct-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
Kontext?Groesse des Kontextfensters in Token. 0 = der beim Training verwendete Kontext des Modells.49152
np12
execution_typelocal

All benchmarks of this model To leaderboard

Anzeige
Videoreihe und Benchmarkauswertung zur AMD R9700
Model comparison

Qwen2.5-3B-Instruct 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

4603452301150,005.83111.66217.494Prefill (tok/s)Generation (tok/s)AMD Radeon RX 7900 XTX - 358,2 tok/s Generation, 14.238 tok/s Prefill, TTFT 1.134 ms (9 Laufe)AMD Radeon RX 7900 XTXNVIDIA GeForce RTX 2060 - 259,9 tok/s Generation, 6.911 tok/s Prefill, TTFT 1.971 ms (6 Laufe)NVIDIA GeForce RTX 20...CPU-only - 66,0 tok/s Generation, 895 tok/s Prefill, TTFT 166.963 ms (5 Laufe)CPU-onlyAMD Radeon AI PRO R9700 - 364,1 tok/s Generation, 11.369 tok/s Prefill, TTFT 1.151 ms (3 Laufe) | DIESER LAUF★ AMD Radeon AI PRO R97...
★ AMD Radeon AI PRO R9700 364,1 tok/s this runAMD Radeon RX 7900 XTX 358,2 tok/sNVIDIA GeForce RTX 2060 259,9 tok/sCPU-only 66,0 tok/s

CPUby processor

4733552361180,006.47312.94619.419Prefill (tok/s)Generation (tok/s)AMD Ryzen Threadripper PRO 3955WX 16-Cores - 358,2 tok/s Generation, 15.665 tok/s Prefill, TTFT 1.104 ms (6 Laufe)AMD Ryzen Threadrippe...Intel(R) Xeon(R) CPU E5-4657L v2 @ 2.40GHz - 333,2 tok/s Generation, 8.401 tok/s Prefill, TTFT 1.712 ms (9 Laufe)Intel(R) Xeon(R) CPU ...Apple M6 (unified CPU/GPU SoC, ARM64, AdvSIMD) - 66,0 tok/s Generation, 2.669 tok/s Prefill, TTFT 988 ms (1 Lauf)Apple M6 (unified CPU...Apple M6 (unified CPU/GPU SoC) - 57,5 tok/s Generation, 1.718 tok/s Prefill, TTFT 1.461 ms (1 Lauf)Apple M6 (unified CPU...4 x Cortex-A78C + 4 x Cortex-A55 - 3,5 tok/s Generation, 30 tok/s Prefill, TTFT 277.455 ms (3 Laufe)4 x Cortex-A78C + 4 x...AMD Ryzen 3 3100 4-Core Processor - 364,1 tok/s Generation, 11.369 tok/s Prefill, TTFT 1.151 ms (3 Laufe) | DIESER LAUF★ AMD Ryzen 3 3100 4-Co...
★ AMD Ryzen 3 3100 4-Core Processor 364,1 tok/s this runAMD Ryzen Threadripper PRO 3955WX 16-Cores 358,2 tok/sIntel(R) Xeon(R) CPU E5-4657L v2 @ 2.40GHz 333,2 tok/sApple M6 (unified CPU/GPU SoC, ARM64, AdvSIMD) 66,0 tok/sApple M6 (unified CPU/GPU SoC) 57,5 tok/s4 x Cortex-A78C + 4 x Cortex-A55 3,5 tok/s

MBby mainboard

4733552361180,006.47312.94619.419Prefill (tok/s)Generation (tok/s)ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI - 358,2 tok/s Generation, 15.665 tok/s Prefill, TTFT 1.104 ms (6 Laufe)ASUSTeK COMPUTER INC....Dell Inc. PowerEdge R820 - 333,2 tok/s Generation, 8.401 tok/s Prefill, TTFT 1.712 ms (9 Laufe)Dell Inc. PowerEdge R...Apple Mac mini - 66,0 tok/s Generation, 2.193 tok/s Prefill, TTFT 1.224 ms (2 Laufe)Apple Mac miniunbekannt - 3,5 tok/s Generation, 30 tok/s Prefill, TTFT 277.455 ms (3 Laufe)unbekanntASUSTeK COMPUTER INC. ROG STRIX X570-F GAMING - 364,1 tok/s Generation, 11.369 tok/s Prefill, TTFT 1.151 ms (3 Laufe) | DIESER LAUF★ ASUSTeK COMPUTER INC....
★ ASUSTeK COMPUTER INC. ROG STRIX X570-F GAMING 364,1 tok/s this runASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 358,2 tok/sDell Inc. PowerEdge R820 333,2 tok/sApple Mac mini 66,0 tok/sunbekannt 3,5 tok/s

ENGby engine

4623462311150,007.05114.10221.152Prefill (tok/s)Generation (tok/s)vLLM - 355,9 tok/s Generation, 17.308 tok/s Prefill, TTFT 1.253 ms (3 Laufe)vLLMlmstudio-mlx - 66,0 tok/s Generation, 2.669 tok/s Prefill, TTFT 988 ms (1 Lauf)lmstudio-mlxOllama - 57,5 tok/s Generation, 1.718 tok/s Prefill, TTFT 1.461 ms (1 Lauf)Ollamallama.cpp - 364,1 tok/s Generation, 8.438 tok/s Prefill, TTFT 47.449 ms (18 Laufe) | DIESER LAUF★ llama.cpp
★ llama.cpp 364,1 tok/s this runvLLM 355,9 tok/slmstudio-mlx 66,0 tok/sOllama 57,5 tok/s

DRVby driver

4013823643463288.5088.8709.2339.595Prefill (tok/s)Generation (tok/s)unbekannt - 364,1 tok/s Generation, 9.051 tok/s Prefill, TTFT 37.404 ms (23 Laufe)unbekannt
unbekannt 364,1 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 20 W
⚡ TDP 300 W
☁️ External LLM (API)
Electricity0.30 EUR/kWh
Avg power (incl. idle)300 W estimated (TDP)GPU 300 W full load
Avg cost / hourEUR 0.090
Electricity / 1M tokensEUR 0.086
Token / kWh3.47M
Acquisition (system)EUR 1,520 partial priceGPU EUR 1,400 · PSU EUR 120
Electricity (2 years)–
TCO (2 years)EUR 3,097
Output tokens (2 years)18.25B
☁️ 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 (20 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

Qwen2.5-3B-InstructAMD Radeon AI PRO R9700Qwen2.5-3B-Instruct2x AMD Radeon RX 7900 XTXQwen2.5-3B-Instruct2x AMD Radeon RX 7900 XTXQwen2.5-3B-Instruct2x AMD Radeon RX 7900 XTX
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.