vL
Modellserver / Inferenz-Engine
vLLM
Alle Performance-Benchmarks, die mit diesem Modellserver gemessen wurden – Durchsatz, Prefill und Latenz.
Beste Generation
1.608,5 tok/s
gemma-4-E2B-it
∅ Generation
265,9 tok/s
Schnitt aus 739 Läufen
Bester Prefill
55.260 tok/s
Prompt-Verarbeitung
Bester TTFT
91 ms
Time to First Token
Top-Modelle auf vLLM
Bester Durchsatz je Modell – umschaltbar nach Generation, Prefill oder Kombiwert.
Auf welcher Hardware laeuft vLLM am schnellsten?
Bester Durchsatz je Grafikkarte mit diesem Modellserver.
Performanceprofil
Durchsatz mit vLLM
Jede Blase steht für eine CPU, GPU bzw. ein Modell – Position: Prompt-Verarbeitung (X) × Ausgabe (Y), Blasengröße: Anzahl Messläufe.
CPUnach Prozessor
AMD Ryzen Threadripper PRO 5975WX 32-Cores 1.608,5 tok/sAMD Ryzen Threadripper PRO 7955WX 16-Cores 1.568,0 tok/sAMD Ryzen 9 9950X 16-Core Processor 1.443,9 tok/sAMD Ryzen 7 5800X3D 8-Core Processor 1.386,8 tok/sAMD Ryzen Threadripper PRO 9965WX 24-Cores 1.238,3 tok/sAMD Ryzen 9 8945HX with Radeon Graphics 901,9 tok/s
GPUnach Grafikkarte
MODnach Modell
gemma-4-E2B-it 1.608,5 tok/sNemotron-3-Nano-4B 1.485,1 tok/sgpt-oss-20b 1.443,9 tok/sQwen3-Coder-30B-A3B-Instruct 1.070,6 tok/sOrnith-1.0-9B 1.062,5 tok/sQwen3.5-35B-A3B 1.028,6 tok/sQwen3-VL-30B-A3B-Instruct 1.028,5 tok/sMeta-Llama-3.1-8B-Instruct 1.026,4 tok/sgemma-4-E4B-it 1.022,8 tok/sMinistral-3-14B-Reasoning-2512 1.015,7 tok/sQwen3-30B-A3B-Instruct-2507 961,5 tok/sQwen3.6-35B-A3B 952,8 tok/sQwen3-Omni-30B-A3B-Thinking 910,7 tok/sQwen3-30B-A3B-Thinking-2507 906,0 tok/sGemma-4-26B-A4B-it 841,8 tok/sLaguna-XS-2.1 836,4 tok/sQwen-AgentWorld-35B-A3B 799,7 tok/sMagistral-Small-2509 726,9 tok/sDevstral-Small-2-24B-Instruct-2512 725,6 tok/sgemma-4-12B-it 681,8 tok/sNorth-Mini-Code-1.0 648,5 tok/sQwen3-Coder-Next 629,8 tok/sOrnith-1.5-35B-A3B 626,4 tok/s
Alle Messungen
← Alle Engines80 Benchmark-Läufe
Jeder veroeffentlichte Lauf mit vLLM.
| # | Modell / Hersteller | Messwerte | Parallel | GPU / CPU / RAM | Runtime | ||
|---|---|---|---|---|---|---|---|
| 1 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.608,46 tok/s TG Prefill 31.452 · TTFT 1.633 ms | 10× | NVIDIA GeForce RTX 5070 TiAMD Ryzen Threadripper PRO 5975WX 32-Cores · 247 GB RAM | vLLMgodclaw | Details → | |
| 2 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.567,98 tok/s TG Prefill 55.260 · TTFT 492 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 3 | Nemotron-3-Nano-4B4BNVIDIA PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.485,14 tok/s TG Prefill 44.171 · TTFT 717 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 4 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.443,88 tok/s TG Prefill 35.348 · TTFT 560 ms | 10× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclaw | Details → | |
| 5 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.386,79 tok/s TG Prefill 19.981 · TTFT 1.205 ms | 10× | NVIDIA GeForce RTX 5090AMD Ryzen 7 5800X3D 8-Core Processor · 126 GB RAM | vLLMcodex_cli | Details → | |
| 6 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.359,09 tok/s TG Prefill 47.158 · TTFT 587 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 7 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.295,30 tok/s TG Prefill 25.462 · TTFT 808 ms | 10× | NVIDIA GeForce RTX 5090AMD Ryzen 7 5800X3D 8-Core Processor · 126 GB RAM | vLLMcodex_cliAWQ | Details → | |
| 8 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.238,26 tok/s TG Prefill 21.843 · TTFT 905 ms | 10× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | vLLMgodclaw | Details → | |
| 9 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.122,50 tok/s TG Prefill 38.712 · TTFT 690 ms | 10× | NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMopenclaw_cliFP8 | Details → | |
| 10 | Qwen3-Coder-30B-A3B-Instruct30BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.070,56 tok/s TG Prefill 33.741 · TTFT 976 ms | 10× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 11 | Ornith-1.0-9B9BDeepReinforce PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.062,52 tok/s TG Prefill 21.965 · TTFT 1.226 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 12 | Qwen3.5-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.028,59 tok/s TG Prefill 24.461 · TTFT 1.670 ms | 10× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 13 | Qwen3-VL-30B-A3B-Instruct30BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.028,53 tok/s TG Prefill 36.884 · TTFT 916 ms | 10× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 14 | Meta-Llama-3.1-8B-Instruct8BMeta PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.026,44 tok/s TG Prefill 26.805 · TTFT 1.371 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 15 | gemma-4-E4B-it8BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.022,79 tok/s TG Prefill 18.744 · TTFT 2.204 ms | 10× | NVIDIA GeForce RTX 5070 TiAMD Ryzen Threadripper PRO 5975WX 32-Cores · 247 GB RAM | vLLMgodclaw | Details → | |
| 16 | Ministral-3-14B-Reasoning-251214BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.015,69 tok/s TG Prefill 20.061 · TTFT 1.289 ms | 10× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 17 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 973,54 tok/s TG Prefill 23.599 · TTFT 460 ms | 5× | NVIDIA GeForce RTX 5070 TiAMD Ryzen Threadripper PRO 5975WX 32-Cores · 247 GB RAM | vLLMgodclaw | Details → | |
| 18 | Qwen3-Coder-30B-A3B-Instruct30BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 966,21 tok/s TG Prefill 33.768 · TTFT 1.793 ms | 10× | NVIDIA GeForce RTX 5090AMD Ryzen 7 5800X3D 8-Core Processor · 126 GB RAM | vLLMcodex_cliAWQ | Details → | |
| 19 | Qwen3-30B-A3B-Instruct-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 961,53 tok/s TG Prefill 32.305 · TTFT 1.633 ms | 10× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 20 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 952,78 tok/s TG Prefill 20.106 · TTFT 1.542 ms | 10× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 21 | Qwen3-Omni-30B-A3B-Thinking30BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 910,74 tok/s TG Prefill 33.907 · TTFT 1.527 ms | 10× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 22 | Qwen3-30B-A3B-Instruct-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 909,25 tok/s TG Prefill 37.116 · TTFT 1.783 ms | 10× | NVIDIA GeForce RTX 5090AMD Ryzen 7 5800X3D 8-Core Processor · 126 GB RAM | vLLMcodex_cliAWQ | Details → | |
| 23 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 909,00 tok/s TG Prefill 43.093 · TTFT 288 ms | 5× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 24 | Qwen3-30B-A3B-Thinking-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 906,04 tok/s TG Prefill 39.029 · TTFT 1.193 ms | 10× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 25 | Qwen3-30B-A3B-Instruct-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 901,86 tok/s TG Prefill 12.965 · TTFT 12.663 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 9 8945HX with Radeon Graphics · 92 GB RAM | vLLMwebsocketAWQ | Details → | |
| 26 | Nemotron-3-Nano-4B4BNVIDIA PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 899,41 tok/s TG Prefill 30.214 · TTFT 473 ms | 5× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 27 | Qwen3-30B-A3B-Thinking-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 883,71 tok/s TG Prefill 12.981 · TTFT 12.917 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 9 8945HX with Radeon Graphics · 92 GB RAM | vLLMwebsocketAWQ | Details → | |
| 28 | Ornith-1.0-9B9BDeepReinforce PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 881,75 tok/s TG Prefill 22.565 · TTFT 1.184 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 29 | Qwen3-30B-A3B-Thinking-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 873,72 tok/s TG Prefill 33.023 · TTFT 1.835 ms | 10× | NVIDIA GeForce RTX 5090AMD Ryzen 7 5800X3D 8-Core Processor · 126 GB RAM | vLLMcodex_cliAWQ | Details → | |
| 30 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 842,94 tok/s TG Prefill 31.098 · TTFT 319 ms | 5× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclaw | Details → | |
| 31 | Gemma-4-26B-A4B-it26BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 841,75 tok/s TG Prefill 32.027 · TTFT 923 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 32 | Laguna-XS-2.1Poolside PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 836,37 tok/s TG Prefill 39.157 · TTFT 1.033 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 33 | Meta-Llama-3.1-8B-Instruct8BMeta PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 832,96 tok/s TG Prefill 27.508 · TTFT 1.327 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 34 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 803,96 tok/s TG Prefill 16.236 · TTFT 616 ms | 5× | NVIDIA GeForce RTX 5090AMD Ryzen 7 5800X3D 8-Core Processor · 126 GB RAM | vLLMcodex_cli | Details → | |
| 35 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 803,18 tok/s TG Prefill 35.757 · TTFT 347 ms | 5× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 36 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 800,39 tok/s TG Prefill 32.313 · TTFT 827 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 37 | Qwen-AgentWorld-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 799,65 tok/s TG Prefill 31.954 · TTFT 834 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 38 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 797,18 tok/s TG Prefill 11.820 · TTFT 1.789 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 9 8945HX with Radeon Graphics · 92 GB RAM | vLLMwebsocket | Details → | |
| 39 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 796,82 tok/s TG Prefill 11.892 · TTFT 1.781 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 9 8945HX with Radeon Graphics · 92 GB RAM | vLLMwebsocket | Details → | |
| 40 | Nemotron-3-Nano-4B4BNVIDIA PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 791,54 tok/s TG Prefill 26.710 · TTFT 1.226 ms | 10× | NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMopenclaw_cliFP8 | Details → | |
| 41 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 790,09 tok/s TG Prefill 13.693 · TTFT 1.142 ms | 5× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | vLLMgodclaw | Details → | |
| 42 | Qwen3-Coder-30B-A3B-Instruct30BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 789,24 tok/s TG Prefill 37.676 · TTFT 952 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 43 | Qwen3.5-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 787,25 tok/s TG Prefill 32.522 · TTFT 826 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 44 | Ornith-1.0-9B9BDeepReinforce PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 786,19 tok/s TG Prefill 19.249 · TTFT 1.376 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 45 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 784,45 tok/s TG Prefill 17.724 · TTFT 560 ms | 5× | NVIDIA GeForce RTX 5090AMD Ryzen 7 5800X3D 8-Core Processor · 126 GB RAM | vLLMcodex_cliAWQ | Details → | |
| 46 | Laguna-XS-2.1Poolside PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 771,80 tok/s TG Prefill 34.194 · TTFT 1.184 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 47 | Qwen3-VL-30B-A3B-Instruct30BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 763,50 tok/s TG Prefill 35.900 · TTFT 987 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 48 | Qwen3-30B-A3B-Instruct-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 758,69 tok/s TG Prefill 34.913 · TTFT 992 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 49 | Gemma-4-26B-A4B-it26BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 754,69 tok/s TG Prefill 26.306 · TTFT 1.120 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 50 | Qwen3-30B-A3B-Thinking-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 751,67 tok/s TG Prefill 32.340 · TTFT 1.271 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 51 | Ministral-3-14B-Reasoning-251214BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 735,14 tok/s TG Prefill 16.886 · TTFT 1.945 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 52 | Qwen-AgentWorld-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 732,27 tok/s TG Prefill 28.221 · TTFT 955 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 53 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 729,28 tok/s TG Prefill 27.628 · TTFT 966 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 54 | Magistral-Small-250924BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 726,94 tok/s TG Prefill 10.204 · TTFT 2.619 ms | 10× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 55 | Devstral-Small-2-24B-Instruct-251224BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 725,58 tok/s TG Prefill 11.872 · TTFT 2.430 ms | 10× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 56 | Qwen3-Coder-30B-A3B-Instruct30BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 720,92 tok/s TG Prefill 30.926 · TTFT 1.138 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 57 | Qwen3.5-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 718,47 tok/s TG Prefill 28.493 · TTFT 951 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 58 | Laguna-XS-2.1Poolside PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 712,04 tok/s TG Prefill 35.143 · TTFT 1.137 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 59 | Qwen3-VL-30B-A3B-Instruct30BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 697,04 tok/s TG Prefill 30.227 · TTFT 1.164 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 60 | Qwen3-30B-A3B-Instruct-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 691,32 tok/s TG Prefill 30.805 · TTFT 1.147 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 61 | Gemma-4-26B-A4B-it26BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 690,88 tok/s TG Prefill 29.513 · TTFT 1.000 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 62 | gemma-4-12B-it12BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 681,82 tok/s TG Prefill 16.224 · TTFT 1.899 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 63 | Qwen3-Coder-30B-A3B-Instruct30BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 676,76 tok/s TG Prefill 21.675 · TTFT 725 ms | 5× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 64 | Meta-Llama-3.1-8B-Instruct8BMeta PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 675,29 tok/s TG Prefill 20.479 · TTFT 781 ms | 5× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 65 | Qwen3-30B-A3B-Thinking-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 673,19 tok/s TG Prefill 30.512 · TTFT 1.156 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 66 | Qwen3-Coder-30B-A3B-Instruct30BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 670,65 tok/s TG Prefill 17.357 · TTFT 3.808 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 9 8945HX with Radeon Graphics · 92 GB RAM | vLLMwebsocketAWQ | Details → | |
| 67 | Qwen3-VL-30B-A3B-Instruct30BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 666,52 tok/s TG Prefill 27.804 · TTFT 465 ms | 5× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 68 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 664,59 tok/s TG Prefill 27.704 · TTFT 965 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 69 | Qwen3.5-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 658,68 tok/s TG Prefill 27.910 · TTFT 964 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 70 | Qwen3-30B-A3B-Instruct-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 655,83 tok/s TG Prefill 17.990 · TTFT 931 ms | 5× | NVIDIA GeForce RTX 5090AMD Ryzen 7 5800X3D 8-Core Processor · 126 GB RAM | vLLMcodex_cliAWQ | Details → | |
| 71 | Laguna-XS-2.1Poolside PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 655,42 tok/s TG Prefill 30.671 · TTFT 1.283 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 72 | Ministral-3-14B-Reasoning-251214BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 655,08 tok/s TG Prefill 14.863 · TTFT 2.211 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 73 | Qwen3-30B-A3B-Instruct-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 649,04 tok/s TG Prefill 17.868 · TTFT 764 ms | 5× | NVIDIA RTX PRO 6000 Blackwell Workstation EditionAMD Ryzen 9 9950X 16-Core Processor · 92 GB RAM | vLLMgodclawAWQ | Details → | |
| 74 | North-Mini-Code-1.0Cohere PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 648,52 tok/s TG Prefill 29.491 · TTFT 1.310 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 75 | Ornith-1.0-9B9BDeepReinforce PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 647,82 tok/s TG Prefill 16.265 · TTFT 725 ms | 5× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 76 | Qwen3-30B-A3B-Thinking-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 633,41 tok/s TG Prefill 20.292 · TTFT 767 ms | 5× | NVIDIA GeForce RTX 5090AMD Ryzen 7 5800X3D 8-Core Processor · 126 GB RAM | vLLMcodex_cliAWQ | Details → | |
| 77 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 631,69 tok/s TG Prefill 29.259 · TTFT 405 ms | 5× | NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMopenclaw_cliFP8 | Details → | |
| 78 | Gemma-4-26B-A4B-it26BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 630,94 tok/s TG Prefill 24.767 · TTFT 1.191 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → | |
| 79 | Qwen3-Coder-NextQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 629,83 tok/s TG Prefill 26.016 · TTFT 1.347 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawAWQ | Details → | |
| 80 | Ornith-1.5-35B-A3B35BDeepReinforce PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 626,41 tok/s TG Prefill 28.681 · TTFT 928 ms | 10× | 2x NVIDIA RTX A6000AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | vLLMgodclawFP8 | Details → |
