Contributed byMario AlkaQwen (Alibaba) Qwen3.8-Flash-Next
Harness benchmark · measured on 29.09.2026 13:41
Benchmark-ID
run-20260929-114231-848f68Decision-Benchmark High V 1.0MoE180BRuntime: vllmQuantisierung: NVFP4
ℹ
Public note: Decision-Benchmark High V 1.0. Qwen3.8-Flash-Next-Uncensored-NVFP4, mazinb checkpoint, auf godcore-ai-001. Ohne Thinking (enable_thinking=false). Optimierter Format-Prompt v2: Systemvorgabe fuer exakt einen erlaubten Antwortwert; Optionen als Textzeilen, keine JSON-Antwort/Erklaerung. Identisch zum Qwen3.8-27B-Promptvalidierungslauf (110/200), abweichend vom urspruenglichen Prompt. Gleiche 200 eingefrorene Aufgaben, unveraenderte strikte Bewertung. Einzelne RTX PRO 6000 Blackwell Max-Q 96GB, GPU0; anderes Modell auf GPU1 resident. NVFP4, vLLM 0.1.dev20073+g8e685d198, kein MTP konfiguriert. Temperatur0, top_p1, top_k-1, min_p0, presence_penalty0, repetition_penalty1, seed20260929, max_tokens32768. Concurrency1, Timeout600s, keine Wiederholungen/Antwortextraktion. Bereits laufender warmer Dienst, kein Neustart/Cache-Reset; vor Messung nur GET models. 107/200 richtig; API-Fehler 0; Formatfehler 0. Mittel 194.702 ms, Median 180.440 ms, P95 206.041 ms. HTTP-End-to-End-Zeit inklusive erster Testanfrage. Synthetische Varianten aus zehn Aufgabenfamilien; Unterschiede in Prompt und Rechenbudget bei historischen Vergleichen beachten.
Points107,00von 200,00
Points rate53,50%
Tasks passed107von 200
Total duration38,94s
Ranking in the field
24of 32 systems
Harness benchmark · Primary metric: Score rate (%)
This run is better than 26 % of all comparable systems.
Score rate
54 %
-27 % vs Ø 73,3
Tasks passed
107
-14 % vs Ø 124,4
Total duration
38,9 s
-97 % vs Ø 1.133,5
Distribution in the field37 – 100 %
Wie schlägt sich dieser Benchmark mit anderen Modellen?
Qwen3.8-Flash-Next same model2× 2x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20261023-180243-30647c
100 %
Kolibri-1NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20261008-165122-c5dea7
100 %
Qwen3.6-35B-A3B-AWQNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260929-153935-691df9
100 %
Qwen3.6-35B-A3B-AWQNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260929-091748-cbe456
100 %
Qwen3.8-27B-Uncensored2× 2x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260929-091747-e13da0
100 %
Open-Jev-27B-v1.1NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260929-091747-5d8477
100 %
Kolibri-1NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20261008-125923-3f1672
99 %
Qwen3.8-Flash-Next same model2× 2x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20261023-180239-3519b5
98 %
Qwen3.6-35B-A3B-AWQNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260929-112249-dba0f2
97 %
Qwen3.8-Flash-Next same model2× 2x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20261023-180239-2b7b40
96 %
Kolibri-1NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20261008-125923-43651d
93 %
Kolibri-1NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20261008-125923-5b5304
91 %
Gemma-4-26B-A4B-itNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260722-161611-05d866
91 %
minimax-m2.5-awq3× 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260722-161611-f6cfe1
81 %
Qwen3.8-Flash-Next this run2× 2x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · run-20260929-114231-848f68
54 %
Hardware
GPU: 2x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition · 96 GB VRAM
CPU: AMD Ryzen Threadripper PRO 9965WX 24-Cores
RAM: 125 GB
Mainboard: ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE
Setup
Runtime: vllm
Quantization: NVFP4
Model: Qwen3.8-Flash-Next
Aufgaben (200)
| Task | Status | Points | Duration |
|---|---|---|---|
| 001 - Beschaffung_Optimierung | failed | 0 / 1 | 0,69 s |
| 002 - permission_deny_override | passed | 1 / 1 | 0,18 s |
| 003 - Rechnung_Rundung | passed | 1 / 1 | 0,19 s |
| 004 - Rechnung_Rundung | passed | 1 / 1 | 0,17 s |
| 005 - graph_reachability_blocks | passed | 1 / 1 | 0,17 s |
| 006 - Beschaffung_Optimierung | passed | 1 / 1 | 0,17 s |
| 007 - time_interval_conflicts | passed | 1 / 1 | 0,16 s |
| 008 - time_interval_conflicts | passed | 1 / 1 | 0,17 s |
| 009 - Rechnung_Rundung | failed | 0 / 1 | 0,18 s |
| 010 - permission_deny_override | passed | 1 / 1 | 0,17 s |
| 011 - Abstimmung_Quorum | passed | 1 / 1 | 0,17 s |
| 012 - permission_deny_override | passed | 1 / 1 | 0,16 s |
| 013 - permission_deny_override | passed | 1 / 1 | 0,17 s |
| 014 - budget_dependency_selection | passed | 1 / 1 | 0,16 s |
| 015 - graph_reachability_blocks | passed | 1 / 1 | 0,17 s |
| 016 - Buchungen_Korrekturen | failed | 0 / 1 | 0,16 s |
| 017 - budget_dependency_selection | passed | 1 / 1 | 0,16 s |
| 018 - Lager_Haltbarkeit | failed | 0 / 1 | 0,19 s |
| 019 - version_event_reconstruction | failed | 0 / 1 | 0,16 s |
| 020 - permission_deny_override | passed | 1 / 1 | 0,17 s |
| 021 - Abstimmung_Quorum | passed | 1 / 1 | 0,15 s |
| 022 - Rechnung_Rundung | failed | 0 / 1 | 0,15 s |
| 023 - permission_deny_override | passed | 1 / 1 | 0,18 s |
| 024 - version_event_reconstruction | passed | 1 / 1 | 0,16 s |
| 025 - Lager_Haltbarkeit | passed | 1 / 1 | 0,16 s |
| 026 - time_interval_conflicts | passed | 1 / 1 | 0,17 s |
| 027 - permission_deny_override | failed | 0 / 1 | 0,18 s |
| 028 - Abstimmung_Quorum | passed | 1 / 1 | 0,18 s |
| 029 - Beschaffung_Optimierung | passed | 1 / 1 | 0,15 s |
| 030 - version_event_reconstruction | failed | 0 / 1 | 0,17 s |
| 031 - Lager_Haltbarkeit | passed | 1 / 1 | 0,18 s |
| 032 - Beschaffung_Optimierung | failed | 0 / 1 | 0,18 s |
| 033 - budget_dependency_selection | failed | 0 / 1 | 0,17 s |
| 034 - time_interval_conflicts | passed | 1 / 1 | 0,15 s |
| 035 - time_interval_conflicts | passed | 1 / 1 | 0,17 s |
| 036 - Rechnung_Rundung | failed | 0 / 1 | 0,16 s |
| 037 - budget_dependency_selection | failed | 0 / 1 | 0,18 s |
| 038 - time_interval_conflicts | passed | 1 / 1 | 0,18 s |
| 039 - Buchungen_Korrekturen | failed | 0 / 1 | 0,17 s |
| 040 - version_event_reconstruction | failed | 0 / 1 | 0,16 s |
| 041 - time_interval_conflicts | failed | 0 / 1 | 0,17 s |
| 042 - Lager_Haltbarkeit | failed | 0 / 1 | 0,15 s |
| 043 - Rechnung_Rundung | passed | 1 / 1 | 0,15 s |
| 044 - Rechnung_Rundung | failed | 0 / 1 | 0,17 s |
| 045 - Rechnung_Rundung | passed | 1 / 1 | 0,17 s |
| 046 - time_interval_conflicts | failed | 0 / 1 | 0,17 s |
| 047 - version_event_reconstruction | failed | 0 / 1 | 1,91 s |
| 048 - Lager_Haltbarkeit | failed | 0 / 1 | 0,17 s |
| 049 - Lager_Haltbarkeit | failed | 0 / 1 | 0,18 s |
| 050 - Beschaffung_Optimierung | passed | 1 / 1 | 0,18 s |
| 051 - Rechnung_Rundung | passed | 1 / 1 | 0,16 s |
| 052 - Lager_Haltbarkeit | failed | 0 / 1 | 0,16 s |
| 053 - Buchungen_Korrekturen | passed | 1 / 1 | 0,17 s |
| 054 - budget_dependency_selection | passed | 1 / 1 | 0,14 s |
| 055 - Buchungen_Korrekturen | passed | 1 / 1 | 0,16 s |
| 056 - Abstimmung_Quorum | passed | 1 / 1 | 0,17 s |
| 057 - Buchungen_Korrekturen | failed | 0 / 1 | 0,18 s |
| 058 - Buchungen_Korrekturen | passed | 1 / 1 | 0,15 s |
| 059 - Lager_Haltbarkeit | passed | 1 / 1 | 0,16 s |
| 060 - Rechnung_Rundung | failed | 0 / 1 | 0,16 s |
| 061 - Abstimmung_Quorum | passed | 1 / 1 | 0,17 s |
| 062 - Buchungen_Korrekturen | passed | 1 / 1 | 0,17 s |
| 063 - version_event_reconstruction | passed | 1 / 1 | 0,17 s |
| 064 - Abstimmung_Quorum | passed | 1 / 1 | 0,17 s |
| 065 - permission_deny_override | failed | 0 / 1 | 0,17 s |
| 066 - time_interval_conflicts | passed | 1 / 1 | 0,17 s |
| 067 - Abstimmung_Quorum | passed | 1 / 1 | 0,17 s |
| 068 - Buchungen_Korrekturen | passed | 1 / 1 | 0,17 s |
| 069 - Abstimmung_Quorum | passed | 1 / 1 | 0,17 s |
| 070 - permission_deny_override | passed | 1 / 1 | 0,17 s |
| 071 - permission_deny_override | failed | 0 / 1 | 0,17 s |
| 072 - permission_deny_override | passed | 1 / 1 | 0,17 s |
| 073 - Rechnung_Rundung | failed | 0 / 1 | 0,18 s |
| 074 - version_event_reconstruction | failed | 0 / 1 | 0,18 s |
| 075 - Buchungen_Korrekturen | passed | 1 / 1 | 0,17 s |
| 076 - graph_reachability_blocks | passed | 1 / 1 | 0,17 s |
| 077 - graph_reachability_blocks | passed | 1 / 1 | 0,17 s |
| 078 - time_interval_conflicts | failed | 0 / 1 | 0,15 s |
| 079 - time_interval_conflicts | passed | 1 / 1 | 0,17 s |
| 080 - graph_reachability_blocks | failed | 0 / 1 | 0,16 s |
| 081 - budget_dependency_selection | failed | 0 / 1 | 0,16 s |
| 082 - version_event_reconstruction | failed | 0 / 1 | 0,17 s |
| 083 - Beschaffung_Optimierung | passed | 1 / 1 | 0,20 s |
| 084 - Beschaffung_Optimierung | passed | 1 / 1 | 0,16 s |
| 085 - permission_deny_override | passed | 1 / 1 | 0,19 s |
| 086 - Lager_Haltbarkeit | failed | 0 / 1 | 0,18 s |
| 087 - budget_dependency_selection | failed | 0 / 1 | 0,18 s |
| 088 - Beschaffung_Optimierung | failed | 0 / 1 | 0,17 s |
| 089 - Rechnung_Rundung | passed | 1 / 1 | 0,19 s |
| 090 - Abstimmung_Quorum | passed | 1 / 1 | 0,18 s |
| 091 - Rechnung_Rundung | failed | 0 / 1 | 0,16 s |
| 092 - graph_reachability_blocks | failed | 0 / 1 | 0,16 s |
| 093 - Buchungen_Korrekturen | failed | 0 / 1 | 0,16 s |
| 094 - Buchungen_Korrekturen | passed | 1 / 1 | 0,15 s |
| 095 - version_event_reconstruction | passed | 1 / 1 | 0,19 s |
| 096 - Buchungen_Korrekturen | passed | 1 / 1 | 0,21 s |
| 097 - graph_reachability_blocks | failed | 0 / 1 | 0,19 s |
| 098 - Buchungen_Korrekturen | passed | 1 / 1 | 0,20 s |
| 099 - budget_dependency_selection | passed | 1 / 1 | 0,20 s |
| 100 - Abstimmung_Quorum | passed | 1 / 1 | 0,18 s |
| 101 - permission_deny_override | passed | 1 / 1 | 0,20 s |
| 102 - Beschaffung_Optimierung | failed | 0 / 1 | 0,19 s |
| 103 - Lager_Haltbarkeit | passed | 1 / 1 | 0,18 s |
| 104 - Lager_Haltbarkeit | passed | 1 / 1 | 0,19 s |
| 105 - time_interval_conflicts | passed | 1 / 1 | 0,19 s |
| 106 - graph_reachability_blocks | failed | 0 / 1 | 0,19 s |
| 107 - version_event_reconstruction | failed | 0 / 1 | 0,20 s |
| 108 - Abstimmung_Quorum | passed | 1 / 1 | 0,18 s |
| 109 - permission_deny_override | passed | 1 / 1 | 0,20 s |
| 110 - Abstimmung_Quorum | passed | 1 / 1 | 0,18 s |
| 111 - Abstimmung_Quorum | passed | 1 / 1 | 0,21 s |
| 112 - Lager_Haltbarkeit | failed | 0 / 1 | 0,20 s |
| 113 - permission_deny_override | failed | 0 / 1 | 0,20 s |
| 114 - version_event_reconstruction | failed | 0 / 1 | 0,19 s |
| 115 - version_event_reconstruction | failed | 0 / 1 | 0,21 s |
| 116 - budget_dependency_selection | failed | 0 / 1 | 0,19 s |
| 117 - Rechnung_Rundung | passed | 1 / 1 | 0,19 s |
| 118 - Buchungen_Korrekturen | passed | 1 / 1 | 0,18 s |
| 119 - Rechnung_Rundung | passed | 1 / 1 | 0,19 s |
| 120 - graph_reachability_blocks | failed | 0 / 1 | 0,50 s |
| 121 - version_event_reconstruction | failed | 0 / 1 | 0,21 s |
| 122 - graph_reachability_blocks | failed | 0 / 1 | 0,18 s |
| 123 - graph_reachability_blocks | failed | 0 / 1 | 0,19 s |
| 124 - permission_deny_override | failed | 0 / 1 | 0,19 s |
| 125 - Lager_Haltbarkeit | passed | 1 / 1 | 0,20 s |
| 126 - budget_dependency_selection | failed | 0 / 1 | 0,19 s |
| 127 - Lager_Haltbarkeit | passed | 1 / 1 | 0,18 s |
| 128 - Rechnung_Rundung | failed | 0 / 1 | 0,19 s |
| 129 - budget_dependency_selection | passed | 1 / 1 | 0,18 s |
| 130 - time_interval_conflicts | passed | 1 / 1 | 0,17 s |
| 131 - Abstimmung_Quorum | failed | 0 / 1 | 0,45 s |
| 132 - Buchungen_Korrekturen | failed | 0 / 1 | 0,18 s |
| 133 - budget_dependency_selection | failed | 0 / 1 | 0,18 s |
| 134 - Buchungen_Korrekturen | failed | 0 / 1 | 0,20 s |
| 135 - time_interval_conflicts | passed | 1 / 1 | 0,19 s |
| 136 - time_interval_conflicts | passed | 1 / 1 | 0,20 s |
| 137 - Beschaffung_Optimierung | passed | 1 / 1 | 0,19 s |
| 138 - time_interval_conflicts | failed | 0 / 1 | 0,20 s |
| 139 - Beschaffung_Optimierung | failed | 0 / 1 | 0,20 s |
| 140 - time_interval_conflicts | failed | 0 / 1 | 0,19 s |
| 141 - version_event_reconstruction | failed | 0 / 1 | 0,20 s |
| 142 - graph_reachability_blocks | failed | 0 / 1 | 0,18 s |
| 143 - budget_dependency_selection | failed | 0 / 1 | 0,18 s |
| 144 - Buchungen_Korrekturen | passed | 1 / 1 | 0,20 s |
| 145 - version_event_reconstruction | failed | 0 / 1 | 0,21 s |
| 146 - Buchungen_Korrekturen | passed | 1 / 1 | 0,18 s |
| 147 - graph_reachability_blocks | failed | 0 / 1 | 0,19 s |
| 148 - time_interval_conflicts | failed | 0 / 1 | 0,17 s |
| 149 - budget_dependency_selection | failed | 0 / 1 | 0,16 s |
| 150 - Rechnung_Rundung | failed | 0 / 1 | 0,20 s |
| 151 - permission_deny_override | passed | 1 / 1 | 0,21 s |
| 152 - version_event_reconstruction | failed | 0 / 1 | 0,19 s |
| 153 - Rechnung_Rundung | passed | 1 / 1 | 0,18 s |
| 154 - Beschaffung_Optimierung | passed | 1 / 1 | 0,20 s |
| 155 - version_event_reconstruction | failed | 0 / 1 | 0,19 s |
| 156 - Abstimmung_Quorum | passed | 1 / 1 | 0,19 s |
| 157 - Abstimmung_Quorum | passed | 1 / 1 | 0,19 s |
| 158 - Lager_Haltbarkeit | passed | 1 / 1 | 0,18 s |
| 159 - budget_dependency_selection | passed | 1 / 1 | 0,18 s |
| 160 - budget_dependency_selection | failed | 0 / 1 | 0,19 s |
| 161 - time_interval_conflicts | passed | 1 / 1 | 0,19 s |
| 162 - graph_reachability_blocks | passed | 1 / 1 | 0,37 s |
| 163 - graph_reachability_blocks | failed | 0 / 1 | 0,19 s |
| 164 - Abstimmung_Quorum | passed | 1 / 1 | 0,20 s |
| 165 - budget_dependency_selection | failed | 0 / 1 | 0,20 s |
| 166 - Beschaffung_Optimierung | failed | 0 / 1 | 0,20 s |
| 167 - Buchungen_Korrekturen | passed | 1 / 1 | 0,19 s |
| 168 - Beschaffung_Optimierung | failed | 0 / 1 | 0,19 s |
| 169 - Beschaffung_Optimierung | passed | 1 / 1 | 0,19 s |
| 170 - budget_dependency_selection | passed | 1 / 1 | 0,18 s |
| 171 - Lager_Haltbarkeit | failed | 0 / 1 | 0,20 s |
| 172 - Rechnung_Rundung | failed | 0 / 1 | 0,18 s |
| 173 - Abstimmung_Quorum | failed | 0 / 1 | 0,17 s |
| 174 - Beschaffung_Optimierung | failed | 0 / 1 | 0,18 s |
| 175 - Beschaffung_Optimierung | passed | 1 / 1 | 0,17 s |
| 176 - version_event_reconstruction | failed | 0 / 1 | 0,18 s |
| 177 - Beschaffung_Optimierung | failed | 0 / 1 | 0,19 s |
| 178 - Lager_Haltbarkeit | passed | 1 / 1 | 0,19 s |
| 179 - permission_deny_override | passed | 1 / 1 | 0,18 s |
| 180 - permission_deny_override | failed | 0 / 1 | 0,20 s |
| 181 - version_event_reconstruction | failed | 0 / 1 | 0,21 s |
| 182 - graph_reachability_blocks | passed | 1 / 1 | 0,20 s |
| 183 - Lager_Haltbarkeit | failed | 0 / 1 | 0,20 s |
| 184 - time_interval_conflicts | passed | 1 / 1 | 0,19 s |
| 185 - Abstimmung_Quorum | passed | 1 / 1 | 0,20 s |
| 186 - Rechnung_Rundung | failed | 0 / 1 | 0,18 s |
| 187 - Beschaffung_Optimierung | passed | 1 / 1 | 0,19 s |
| 188 - Buchungen_Korrekturen | failed | 0 / 1 | 0,19 s |
| 189 - graph_reachability_blocks | failed | 0 / 1 | 0,19 s |
| 190 - Abstimmung_Quorum | passed | 1 / 1 | 0,20 s |
| 191 - budget_dependency_selection | passed | 1 / 1 | 0,18 s |
| 192 - permission_deny_override | passed | 1 / 1 | 0,19 s |
| 193 - graph_reachability_blocks | passed | 1 / 1 | 0,18 s |
| 194 - budget_dependency_selection | passed | 1 / 1 | 0,19 s |
| 195 - Lager_Haltbarkeit | failed | 0 / 1 | 0,20 s |
| 196 - graph_reachability_blocks | failed | 0 / 1 | 0,18 s |
| 197 - graph_reachability_blocks | failed | 0 / 1 | 0,20 s |
| 198 - Lager_Haltbarkeit | passed | 1 / 1 | 0,18 s |
| 199 - Beschaffung_Optimierung | failed | 0 / 1 | 0,20 s |
| 200 - version_event_reconstruction | failed | 0 / 1 | 0,19 s |
Configuration
# LLM-Benchmark Konfiguration
# Modell : Qwen3.8-Flash-Next
# Engine : vLLM
# Run-ID : run-20260929-114231-848f68
# GPU : 2x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
# CPU : AMD Ryzen Threadripper PRO 9965WX 24-Cores
# RAM : 125 GB
bench@llm-benchmark:~$ vllm serve \
--model Qwen3.8-Flash-Next-Uncensored-NVFP4 \
temperature 0 \
top_p 1 \
top_k -1 \
min_p 0 \
presence_penalty 0 \
repetition_penalty 1 \
seed 20260929 \
max_tokens 32768 \
model Qwen3.8-Flash-Next-Uncensored \
endpoint http://10.220.4.10:8000/v1/chat/completions
| 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. | Qwen3.8-Flash-Next-Uncensored |
| temperature | 0 |
| top_p | 1 |
| top_k | -1 |
| min_p | 0 |
| presence_penalty | 0 |
| repetition_penalty | 1 |
| Seed?Zufalls-Seed fuer reproduzierbare Ausgaben. -1 = zufaellig. | 20260929 |
| max_tokens | 32768 |
| enable_thinking | inaktiv |
| Modellpfad?Pfad zur GGUF-Modelldatei, die geladen und ausgeliefert wird. | Qwen3.8-Flash-Next-Uncensored |
| endpoint | http://10.220.4.10:8000/v1/chat/completions |
| execution_type | local |
Anzeige
Model comparison
Qwen3.8-Flash-Next 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 Max-Q Workstation Edition 1.564,6 tok/s this runNVIDIA RTX A6000 1.171,8 tok/sNVIDIA GeForce RTX 3090 1.132,7 tok/sNVIDIA RTX PRO 6000 Blackwell Workstation Edition 584,1 tok/sNVIDIA DGX SPARK (GB10) 492,8 tok/sNVIDIA GeForce RTX 5090 346,0 tok/sNVIDIA Tesla V100 PCIe 32GB 196,2 tok/sAMD Radeon RX 7900 XTX 18,7 tok/s
CPUby processor
AMD Ryzen 9 7945HX with Radeon Graphics 1.564,6 tok/sAMD Ryzen Threadripper PRO 3945WX 12-Cores 1.171,8 tok/sAMD Ryzen 5 5600X 6-Core Processor 1.132,7 tok/sAMD Ryzen 9 9950X 16-Core Processor 584,1 tok/sNVIDIA Grace 492,8 tok/s★ AMD Ryzen Threadripper PRO 9965WX 24-Cores 424,2 tok/s this runAMD Ryzen 7 5800X3D 8-Core Processor 346,0 tok/sAMD Ryzen 3 PRO 3200GE w/ Radeon Vega Graphics 196,2 tok/sAMD EPYC 7203P 8-Core Processor 79,2 tok/sIntel(R) Xeon(R) CPU E5-4657L v2 @ 2.40GHz 63,2 tok/sAMD Ryzen Threadripper PRO 3955WX 16-Cores 18,7 tok/s
MBby mainboard
Shenzhen Meigao Electronic Equipment Co.,Ltd DRFXI (MotherBoard Series) 1.564,6 tok/sunbekannt 1.171,8 tok/sASUSTeK COMPUTER INC. ROG STRIX B550-A GAMING 1.132,7 tok/sASUSTeK COMPUTER INC. ProArt X870E-CREATOR WIFI 584,1 tok/sASUSTeK COMPUTER INC. GX10 492,8 tok/s★ ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE 424,2 tok/s this runASUSTeK COMPUTER INC. ROG STRIX X570-F GAMING 196,2 tok/sASRockRack ROMED8-2T/BCM 79,2 tok/sDell Inc. PowerEdge R820 63,2 tok/sASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 35,4 tok/s
ENGby engine
DRVby driver
unbekannt 1.564,6 tok/sNVIDIA 590.48.01 / CUDA 13.1 492,8 tok/s
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.