Created bymario-alka.dePowered bygodcore.denoob2claw.detricoma.de
Contributed byMario AlkaQwen (Alibaba)

Qwen2.5-Coder-32B-Instruct

Performance benchmark · measured on 29.07.2026 04:07

Benchmark-IDrun-20260729-060801-4a9c3f
Timebench 3 - Kombi (Prefill + Generation)Dense32BRuntime: llama.cppQuantisierung: Q4_K_M
Generation25,71tok/s
Prefill662,09tok/s
Time to First Token64.391,00ms
Total duration756,19s
Concurrency5parallel
Ranking in the field
44of 59 systems

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

This run is better than 26 % of all comparable systems.
Generation 25,7 tok/s
-87 % vs Ø 192,8
Prefill 662,1 tok/s
-74 % vs Ø 2.593,3
Time to First Token 64.391 ms
+55 % vs Ø 41.662
Distribution in the field4 – 974 tok/s
Ø 193 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 GeForce RTX 5070 Ti · 16 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: Q4_K_M
Model: Qwen2.5-Coder-32B-Instruct

Configuration

benchmark-konfiguration — run-20260729-060801-4a9c3f
# LLM-Benchmark Konfiguration # Modell : Qwen2.5-Coder-32B-Instruct # Engine : llama.cpp # Run-ID : run-20260729-060801-4a9c3f # GPU : NVIDIA GeForce RTX 5070 Ti # CPU : AMD Ryzen Threadripper PRO 5975WX 32-Cores # RAM : 247 GB bench@llm-benchmark:~$ /home/godcore/llama.cpp/build/bin/llama-server \ -m /home/godcore/.cache/huggingface/hub/models--bartowski--Qwen2.5-Coder-32B-Instruct-GGUF/snapshots/40b525506a4f98ed425882fa6dfc90cc8139065e/Qwen2.5-Coder-32B-Instruct-Q4_K_M.gguf \ --alias Qwen2.5-Coder-32B-Instruct \ --host 0.0.0.0 \ --port 8000 \ -ngl 30 \ -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.Qwen2.5-Coder-32B-Instruct
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--bartowski--Qwen2.5-Coder-32B-Instruct-GGUF/snapshots/40b525506a4f98ed425882fa6dfc90cc8139065e/Qwen2.5-Coder-32B-Instruct-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.30
Kontext?Groesse des Kontextfensters in Token. 0 = der beim Training verwendete Kontext des Modells.16384
np4

All benchmarks of this model To leaderboard

Anzeige
Model comparison

Qwen2.5-Coder-32B-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

6474853231620,001.8063.6135.419Prefill (tok/s)Generation (tok/s)NVIDIA RTX PRO 6000 Blackwell Workstation Edition - 500,6 tok/s Generation, 4.446 tok/s Prefill, TTFT 6.330 ms (6 Laufe)NVIDIA RTX PRO 6000 B...AMD Radeon AI PRO R9700 - 63,2 tok/s Generation, 2.588 tok/s Prefill, TTFT 3.127 ms (10 Laufe)AMD Radeon AI PRO R97...AMD Radeon 8060S Graphics - 20,6 tok/s Generation, 518 tok/s Prefill, TTFT 24.112 ms (3 Laufe)AMD Radeon 8060S Grap...NVIDIA GeForce RTX 5070 Ti - 45,3 tok/s Generation, 587 tok/s Prefill, TTFT 95.842 ms (3 Laufe) | DIESER LAUF★ NVIDIA GeForce RTX 50...
NVIDIA RTX PRO 6000 Blackwell Workstation Edition 500,6 tok/sAMD Radeon AI PRO R9700 63,2 tok/s★ NVIDIA GeForce RTX 5070 Ti 45,3 tok/s this runAMD Radeon 8060S Graphics 20,6 tok/s

CPUby processor

6474853231620,001.8063.6135.419Prefill (tok/s)Generation (tok/s)AMD Ryzen 9 9950X 16-Core Processor - 500,6 tok/s Generation, 4.446 tok/s Prefill, TTFT 6.330 ms (6 Laufe)AMD Ryzen 9 9950X 16-...AMD Ryzen Threadripper PRO 7955WX 16-Cores - 63,2 tok/s Generation, 2.588 tok/s Prefill, TTFT 3.127 ms (10 Laufe)AMD Ryzen Threadrippe...AMD RYZEN AI MAX+ 395 w/ Radeon 8060S - 20,6 tok/s Generation, 518 tok/s Prefill, TTFT 24.112 ms (3 Laufe)AMD RYZEN AI MAX+ 395...AMD Ryzen Threadripper PRO 5975WX 32-Cores - 45,3 tok/s Generation, 587 tok/s Prefill, TTFT 95.842 ms (3 Laufe) | DIESER LAUF★ AMD Ryzen Threadrippe...
AMD Ryzen 9 9950X 16-Core Processor 500,6 tok/sAMD Ryzen Threadripper PRO 7955WX 16-Cores 63,2 tok/s★ AMD Ryzen Threadripper PRO 5975WX 32-Cores 45,3 tok/s this runAMD RYZEN AI MAX+ 395 w/ Radeon 8060S 20,6 tok/s

MBby mainboard

6474853231620,001.8063.6135.419Prefill (tok/s)Generation (tok/s)ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI - 500,6 tok/s Generation, 4.446 tok/s Prefill, TTFT 6.330 ms (6 Laufe)ASUSTeK COMPUTER INC....ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE - 63,2 tok/s Generation, 2.588 tok/s Prefill, TTFT 3.127 ms (10 Laufe)ASUSTeK COMPUTER INC....Bosgame AXB35-02 (BeyondMax Series) - 20,6 tok/s Generation, 518 tok/s Prefill, TTFT 24.112 ms (3 Laufe)Bosgame AXB35-02 (Bey...ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI - 45,3 tok/s Generation, 587 tok/s Prefill, TTFT 95.842 ms (3 Laufe) | DIESER LAUF★ ASUSTeK COMPUTER INC....
ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI 500,6 tok/sASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE 63,2 tok/s★ ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 45,3 tok/s this runBosgame AXB35-02 (BeyondMax Series) 20,6 tok/s

ENGby engine

5555224894554228352.5934.3526.110Prefill (tok/s)Generation (tok/s)vLLM - 500,6 tok/s Generation, 5.183 tok/s Prefill, TTFT 1.853 ms (5 Laufe)vLLMllama.cpp - 476,9 tok/s Generation, 1.762 tok/s Prefill, TTFT 24.697 ms (17 Laufe) | DIESER LAUF★ llama.cpp
vLLM 500,6 tok/s★ llama.cpp 476,9 tok/s this run

DRVby driver

5515265014764512.3872.4892.5902.692Prefill (tok/s)Generation (tok/s)unbekannt - 500,6 tok/s Generation, 2.540 tok/s Prefill, TTFT 19.505 ms (22 Laufe)unbekannt
unbekannt 500,6 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 310 W
☁️ External LLM (API)
Electricity0.30 EUR/kWh
Avg power (incl. idle)310 W estimated (TDP)GPU 300 + Board 10 W full load
Avg cost / hourEUR 0.093
Electricity / 1M tokensEUR 1.00
Token / kWh298.57K
Acquisition (system)EUR 2,126 missingRAM EUR 1,976 · PSU EUR 150
Electricity (2 years)
TCO (2 years)EUR 3,755
Output tokens (2 years)1.62B
☁️ 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

Qwen2.5-Coder-32B-InstructNVIDIA GeForce RTX 5070 TiQwen2.5-Coder-32B-InstructNVIDIA RTX PRO 6000 Blackwell Workstation EditionQwen2.5-Coder-32B-InstructNVIDIA RTX PRO 6000 Blackwell Workstation EditionQwen2.5-Coder-32B-Instruct3x AMD Radeon AI PRO R9700
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.