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

Qwen-AgentWorld-35B-A3B

Performance benchmark · measured on 27.08.2026 20:22

Benchmark-IDrun-20260827-182612-35e531
Timebench 3 - Kombi (Prefill + Generation)MoE35BRuntime: llama.cppQuantisierung: Q4_K_M
Generation139,63tok/s
Prefill3.780,96tok/s
Time to First Token510,00ms
Total duration15,69s
Concurrency1parallel
Ranking in the field
47of 178 systems

Performance benchmark · Primary metric: Generation-Speed (tok/s) · 1× concurrent

This run is better than 74 % of all comparable systems.
Generation 139,6 tok/s
+46 % vs Ø 95,7
Prefill 3.781,0 tok/s
-22 % vs Ø 4.832,0
Time to First Token 510 ms
-26 % vs Ø 691
Distribution in the field5 – 195 tok/s
Ø 96 Median Dieser Lauf

Wie schlägt sich dieser Benchmark mit anderen Modellen?

How does this benchmark compare on other GPUs?

Same model on different hardware · 1× concurrent · Generation (tok/s)

Hardware

GPU: 2x 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: llama.cpp
Quantization: Q4_K_M
Model: Qwen-AgentWorld-35B-A3B

Configuration

benchmark-konfiguration — run-20260827-182612-35e531
# LLM-Benchmark Konfiguration # Modell : Qwen-AgentWorld-35B-A3B # Engine : llama.cpp # Run-ID : run-20260827-182612-35e531 # GPU : 2x NVIDIA RTX A6000 # CPU : AMD Ryzen Threadripper PRO 7955WX 16-Cores # RAM : 184 GB bench@llm-benchmark:~$ /home/godcore/llama.cpp/build/bin/llama-server \ -m Qwen-AgentWorld-35B-A3B-UD-Q4_K_M.gguf \ -ngl 999 \ -fa on \ -c 32768 \ -np 8 \ -sm layer '(2x' RTX A6000 '48GB)'
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
Kontextlaenge?Maximale Anzahl Tokens (Eingabe + erzeugte Ausgabe zusammen), die das Modell pro Anfrage verarbeiten kann.32768
Modellpfad?Pfad zur GGUF-Modelldatei, die geladen und ausgeliefert wird.Qwen-AgentWorld-35B-A3B-UD-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.999
faon
Kontext?Groesse des Kontextfensters in Token. 0 = der beim Training verwendete Kontext des Modells.32768
np8
Split-Mode?Verteilung ueber mehrere GPUs: none (nur eine GPU), layer (Layer aufteilen) oder row (Tensoren zeilenweise).layer

All benchmarks of this model To leaderboard

Anzeige
Model comparison

Qwen-AgentWorld-35B-A3B 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

1.9051.4299524760,003.7447.48811.232Prefill (tok/s)Generation (tok/s)NVIDIA RTX PRO 6000 Blackwell Workstation Edition - 1.481,2 tok/s Generation, 6.880 tok/s Prefill, TTFT 3.627 ms (3 Laufe)NVIDIA RTX PRO 6000 B...NVIDIA GeForce RTX 5090 - 1.439,1 tok/s Generation, 5.167 tok/s Prefill, TTFT 4.136 ms (3 Laufe)NVIDIA GeForce RTX 50...NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition - 1.363,4 tok/s Generation, 6.823 tok/s Prefill, TTFT 3.890 ms (9 Laufe)NVIDIA RTX PRO 6000 B...NVIDIA GeForce RTX 3090 Ti - 849,2 tok/s Generation, 3.458 tok/s Prefill, TTFT 6.729 ms (3 Laufe)NVIDIA GeForce RTX 30...AMD Radeon AI PRO R9700 - 192,0 tok/s Generation, 1.696 tok/s Prefill, TTFT 32.302 ms (21 Laufe)AMD Radeon AI PRO R97...NVIDIA GeForce RTX 5070 Ti - 140,1 tok/s Generation, 468 tok/s Prefill, TTFT 43.146 ms (3 Laufe)NVIDIA GeForce RTX 50...AMD Radeon 8060S Graphics - 103,8 tok/s Generation, 1.220 tok/s Prefill, TTFT 8.514 ms (3 Laufe)AMD Radeon 8060S Grap...NVIDIA RTX A6000 - 732,3 tok/s Generation, 9.126 tok/s Prefill, TTFT 2.180 ms (15 Laufe) | DIESER LAUF★ NVIDIA RTX A6000
NVIDIA RTX PRO 6000 Blackwell Workstation Edition 1.481,2 tok/sNVIDIA GeForce RTX 5090 1.439,1 tok/sNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 1.363,4 tok/sNVIDIA GeForce RTX 3090 Ti 849,2 tok/s★ NVIDIA RTX A6000 732,3 tok/s this runAMD Radeon AI PRO R9700 192,0 tok/sNVIDIA GeForce RTX 5070 Ti 140,1 tok/sAMD Radeon 8060S Graphics 103,8 tok/s

CPUby processor

1.9051.4299524760,002.8165.6328.447Prefill (tok/s)Generation (tok/s)AMD Ryzen 9 9950X 16-Core Processor - 1.481,2 tok/s Generation, 6.880 tok/s Prefill, TTFT 3.627 ms (3 Laufe)AMD Ryzen 9 9950X 16-...AMD Ryzen 7 5800X3D 8-Core Processor - 1.439,1 tok/s Generation, 5.167 tok/s Prefill, TTFT 4.136 ms (3 Laufe)AMD Ryzen 7 5800X3D 8...AMD Ryzen Threadripper PRO 9965WX 24-Cores - 1.363,4 tok/s Generation, 6.823 tok/s Prefill, TTFT 3.890 ms (9 Laufe)AMD Ryzen Threadrippe...AMD Ryzen 9 8945HX with Radeon Graphics - 849,2 tok/s Generation, 3.458 tok/s Prefill, TTFT 6.729 ms (3 Laufe)AMD Ryzen 9 8945HX wi...AMD Ryzen Threadripper PRO 5975WX 32-Cores - 140,1 tok/s Generation, 468 tok/s Prefill, TTFT 43.146 ms (3 Laufe)AMD Ryzen Threadrippe...AMD RYZEN AI MAX+ 395 w/ Radeon 8060S - 103,8 tok/s Generation, 1.220 tok/s Prefill, TTFT 8.514 ms (3 Laufe)AMD RYZEN AI MAX+ 395...AMD Ryzen Threadripper PRO 7955WX 16-Cores - 732,3 tok/s Generation, 4.792 tok/s Prefill, TTFT 19.751 ms (36 Laufe) | DIESER LAUF★ AMD Ryzen Threadrippe...
AMD Ryzen 9 9950X 16-Core Processor 1.481,2 tok/sAMD Ryzen 7 5800X3D 8-Core Processor 1.439,1 tok/sAMD Ryzen Threadripper PRO 9965WX 24-Cores 1.363,4 tok/sAMD Ryzen 9 8945HX with Radeon Graphics 849,2 tok/s★ AMD Ryzen Threadripper PRO 7955WX 16-Cores 732,3 tok/s this runAMD Ryzen Threadripper PRO 5975WX 32-Cores 140,1 tok/sAMD RYZEN AI MAX+ 395 w/ Radeon 8060S 103,8 tok/s

MBby mainboard

1.9051.4299524760,002.8165.6328.447Prefill (tok/s)Generation (tok/s)ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI - 1.481,2 tok/s Generation, 6.880 tok/s Prefill, TTFT 3.627 ms (3 Laufe)ASUSTeK COMPUTER INC....ASUSTeK COMPUTER INC. ROG STRIX B550-A GAMING - 1.439,1 tok/s Generation, 5.167 tok/s Prefill, TTFT 4.136 ms (3 Laufe)ASUSTeK COMPUTER INC....Meigao Innovation Technology (Shen Zhen) Co., Ltd DRFXL (MotherBoard Series) - 849,2 tok/s Generation, 3.458 tok/s Prefill, TTFT 6.729 ms (3 Laufe)Meigao Innovation Tec...ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI - 140,1 tok/s Generation, 468 tok/s Prefill, TTFT 43.146 ms (3 Laufe)ASUSTeK COMPUTER INC....Bosgame AXB35-02 (BeyondMax Series) - 103,8 tok/s Generation, 1.220 tok/s Prefill, TTFT 8.514 ms (3 Laufe)Bosgame AXB35-02 (Bey...ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE - 1.363,4 tok/s Generation, 5.198 tok/s Prefill, TTFT 16.579 ms (45 Laufe) | DIESER LAUF★ ASUSTeK COMPUTER INC....
ASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI 1.481,2 tok/sASUSTeK COMPUTER INC. ROG STRIX B550-A GAMING 1.439,1 tok/s★ ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE 1.363,4 tok/s this runMeigao Innovation Technology (Shen Zhen) Co., Ltd DRFXL (MotherBoard Series) 849,2 tok/sASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 140,1 tok/sBosgame AXB35-02 (BeyondMax Series) 103,8 tok/s

ENGby engine

1.9101.4329554770,006.48012.96019.440Prefill (tok/s)Generation (tok/s)vLLM - 732,3 tok/s Generation, 16.030 tok/s Prefill, TTFT 700 ms (6 Laufe)vLLMunbekannt - 78,5 tok/s Generation, 2.425 tok/s Prefill, TTFT 886 ms (3 Laufe)unbekanntllama.cpp - 1.481,2 tok/s Generation, 3.570 tok/s Prefill, TTFT 18.385 ms (51 Laufe) | DIESER LAUF★ llama.cpp
★ llama.cpp 1.481,2 tok/s this runvLLM 732,3 tok/sunbekannt 78,5 tok/s

DRVby driver

1.9101.4329554770,01.6902.9994.3075.616Prefill (tok/s)Generation (tok/s)unbekannt - 1.481,2 tok/s Generation, 4.881 tok/s Prefill, TTFT 16.524 ms (57 Laufe)unbekanntAMD 7.0.0-27-generic - 78,5 tok/s Generation, 2.425 tok/s Prefill, TTFT 886 ms (3 Laufe)AMD 7.0.0-27-generic
unbekannt 1.481,2 tok/sAMD 7.0.0-27-generic 78,5 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 URLReset
⚡ Electricity price EUR/kWh
⚙️ System utilization 100 %
🖥️ Acquisition EUR
🔌 Idle 95 W
⚡ TDP 671 W
☁️ External LLM (API)
Electricity0.30 EUR/kWh
Avg power (incl. idle)671 W estimated (TDP)GPU 600 + CPU 46 + Board 25 W full load
Avg cost / hourEUR 0.20
Electricity / 1M tokensEUR 0.40
Token / kWh749.13K
Acquisition (system)EUR 11,454 full priceGPU EUR 6,000 · CPU EUR 1,399 · Board EUR 1,299 · RAM EUR 2,576 · PSU EUR 180
Electricity (2 years)
TCO (2 years)EUR 14,981
Output tokens (2 years)8.81B
☁️ 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 (95 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

Qwen-AgentWorld-35B-A3B2x NVIDIA RTX A6000Qwen-AgentWorld-35B-A3BNVIDIA GeForce RTX 5090Qwen-AgentWorld-35B-A3BNVIDIA RTX PRO 6000 Blackwell Workstation EditionQwen-AgentWorld-35B-A3B3x 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)

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.