🏆 Rangliste
Highscore
Die besten Laeufe im direkten Ranking - umschaltbar zwischen Performance (Token/s) und Harness (Aufgabenpunkte).
| # | Modell / Hersteller | Grafikkarte | CPU / RAM | Generation | Prefill | TG | TTFT | Parallel | Runtime | |
|---|---|---|---|---|---|---|---|---|---|---|
| 🥇 | gpt-oss-20bOpenAI | NVIDIA RTX PRO 6000 Blackwell Workstation Edition | 32x AMD Ryzen 9 9950X 16-Core Processor 92 GB | 1.443,9 tok/s |
35.348 | 1.443,9 | 560 ms | 10× | vLLMgodclaw | Details → |
| 🥈 | gpt-oss-20bOpenAI | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 1.295,3 tok/s |
25.462 | 1.295,3 | 808 ms | 10× | vLLMcodex_cliAWQ | Details → |
| 🥉 | Ministral-3-14B-Reasoning-2512Mistral AI | NVIDIA RTX PRO 6000 Blackwell Workstation Edition | 32x AMD Ryzen 9 9950X 16-Core Processor 92 GB | 1.015,7 tok/s |
20.061 | 1.015,7 | 1288 ms | 10× | vLLMgodclawAWQ | Details → |
| 4 | Qwen3-Coder-30B-A3B-InstructQwen (Alibaba) | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 966,2 tok/s |
33.768 | 966,2 | 1792 ms | 10× | vLLMcodex_cliAWQ | Details → |
| 5 | Qwen3-30B-A3B-Instruct-2507Qwen (Alibaba) | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 909,3 tok/s |
37.116 | 909,3 | 1783 ms | 10× | vLLMcodex_cliAWQ | Details → |
| 6 | Qwen3-30B-A3B-Thinking-2507Qwen (Alibaba) | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 873,7 tok/s |
33.023 | 873,7 | 1834 ms | 10× | vLLMcodex_cliAWQ | Details → |
| 7 | gpt-oss-20bOpenAI | NVIDIA RTX PRO 6000 Blackwell Workstation Edition | 32x AMD Ryzen 9 9950X 16-Core Processor 92 GB | 842,9 tok/s |
31.098 | 842,9 | 318 ms | 5× | vLLMgodclaw | Details → |
| 8 | gpt-oss-20bOpenAI | NVIDIA GeForce RTX 3090 Ti | 32x AMD Ryzen 9 8945HX with Radeon Graphics 92 GB | 796,8 tok/s |
11.892 | 796,8 | 1780 ms | 10× | vLLMwebsocket | Details → |
| 9 | gpt-oss-20bOpenAI | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 784,5 tok/s |
17.724 | 784,5 | 560 ms | 5× | vLLMcodex_cliAWQ | Details → |
| 10 | Qwen3-Coder-30B-A3B-InstructQwen (Alibaba) | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 688,6 tok/s |
16.192 | 688,6 | 873 ms | 5× | vLLMcodex_cliAWQ | Details → |
| 11 | Qwen3-30B-A3B-Instruct-2507Qwen (Alibaba) | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 655,8 tok/s |
17.990 | 655,8 | 930 ms | 5× | vLLMcodex_cliAWQ | Details → |
| 12 | Qwen3-30B-A3B-Thinking-2507Qwen (Alibaba) | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 633,4 tok/s |
20.292 | 633,4 | 767 ms | 5× | vLLMcodex_cliAWQ | Details → |
| 13 | Ministral-3-14B-Reasoning-2512Mistral AI | NVIDIA RTX PRO 6000 Blackwell Workstation Edition | 32x AMD Ryzen 9 9950X 16-Core Processor 92 GB | 615,8 tok/s |
13.999 | 615,8 | 796 ms | 5× | vLLMgodclawAWQ | Details → |
| 14 | Ministral-3-14B-Reasoning-2512Mistral AI | NVIDIA GeForce RTX 3090 Ti | 32x AMD Ryzen 9 8945HX with Radeon Graphics 92 GB | 524,4 tok/s |
7.751 | 524,4 | 4130 ms | 10× | vLLMwebsocketAWQ | Details → |
| 15 | gpt-oss-20bOpenAI | NVIDIA GeForce RTX 3090 Ti | 32x AMD Ryzen 9 8945HX with Radeon Graphics 92 GB | 483,7 tok/s |
9.270 | 483,7 | 1069 ms | 5× | vLLMwebsocket | Details → |
| 16 | gpt-oss-20bOpenAI | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 340,5 tok/s |
4.258 | 340,5 | 2328 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 17 | Ministral-3-14B-Reasoning-2512Mistral AI | NVIDIA GeForce RTX 3090 Ti | 32x AMD Ryzen 9 8945HX with Radeon Graphics 92 GB | 338,3 tok/s |
5.943 | 338,3 | 2299 ms | 5× | vLLMwebsocketAWQ | Details → |
| 18 | gpt-oss-20bOpenAI | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 331,7 tok/s |
4.763 | 331,7 | 4261 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 19 | North-Mini-Code-1.0Cohere | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 279,0 tok/s |
6.997 | 279,0 | 3294 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 20 | NVIDIA-Nemotron-3-Nano-30B-A3BNVIDIA | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 274,0 tok/s |
5.579 | 274,0 | 3956 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 21 | Nemotron-Cascade-2-30B-A3BNVIDIA | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 270,4 tok/s |
5.684 | 270,4 | 3915 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 22 | Nemotron-3-Nano-Omni-30B-A3B-ReasoningNVIDIA | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 268,7 tok/s |
5.448 | 268,7 | 4050 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 23 | gpt-oss-20bOpenAI | NVIDIA RTX PRO 6000 Blackwell Workstation Edition | 32x AMD Ryzen 9 9950X 16-Core Processor 92 GB | 268,7 tok/s |
17.371 | 268,7 | 113 ms | 1× | vLLMgodclaw | Details → |
| 24 | gpt-oss-20bOpenAI | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 267,0 tok/s |
7.273 | 267,0 | 271 ms | 1× | vLLMcodex_cliAWQ | Details → |
| 25 | NVIDIA-Nemotron-3-Nano-4BNVIDIA | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 261,2 tok/s |
3.150 | 261,2 | 3127 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 26 | Qwen3-Coder-30B-A3B-InstructQwen (Alibaba) | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 243,0 tok/s |
4.297 | 243,0 | 764 ms | 1× | vLLMcodex_cliAWQ | Details → |
| 27 | gpt-oss-120bOpenAI | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 241,4 tok/s |
3.188 | 241,4 | 6381 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 28 | Qwen3-Coder-30B-A3B-InstructQwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 237,2 tok/s |
11.379 | 237,2 | 2588 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 29 | Qwen3-30B-A3B-Instruct-2507Qwen (Alibaba) | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 232,8 tok/s |
8.323 | 232,8 | 311 ms | 1× | vLLMcodex_cliAWQ | Details → |
| 30 | Qwen3-30B-A3B-Thinking-2507Qwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 231,1 tok/s |
3.322 | 231,1 | 3312 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 31 | NVIDIA-Nemotron-3-Nano-30B-A3BNVIDIA | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 229,1 tok/s |
3.641 | 229,1 | 2704 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 32 | Nemotron-Cascade-2-30B-A3BNVIDIA | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 225,0 tok/s |
3.703 | 225,0 | 2680 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 33 | Nemotron-3-Nano-Omni-30B-A3B-ReasoningNVIDIA | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 223,5 tok/s |
3.548 | 223,5 | 2773 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 34 | North-Mini-Code-1.0Cohere | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 223,3 tok/s |
5.459 | 223,3 | 1907 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 35 | Qwen3-30B-A3B-Thinking-2507Qwen (Alibaba) | NVIDIA GeForce RTX 5090 | 16x AMD Ryzen 7 5800X3D 8-Core Processor 126 GB | 221,3 tok/s |
8.379 | 221,3 | 300 ms | 1× | vLLMcodex_cliAWQ | Details → |
| 36 | NVIDIA-Nemotron-3-Nano-4BNVIDIA | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 218,9 tok/s |
4.706 | 218,9 | 4696 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 37 | gpt-oss-120bOpenAI | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 215,3 tok/s |
2.707 | 215,3 | 3658 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 38 | Qwen3.6-35B-A3BQwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 206,9 tok/s |
2.708 | 206,9 | 3698 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 39 | Qwen-AgentWorld-35B-A3BQwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 206,1 tok/s |
2.716 | 206,1 | 3683 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 40 | Qwen3.5-35B-A3BQwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 205,4 tok/s |
2.760 | 205,4 | 3627 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 41 | Qwen3-Coder-30B-A3B-InstructQwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 205,0 tok/s |
8.401 | 205,0 | 1506 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 42 | Qwen3-30B-A3B-Thinking-2507Qwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 197,5 tok/s |
4.440 | 197,5 | 5503 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 43 | Qwen3.5-35B-A3BQwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 197,2 tok/s |
3.650 | 197,2 | 6184 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 44 | Qwen3-30B-A3B-Instruct-2507Qwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 191,9 tok/s |
4.994 | 191,9 | 2502 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 45 | Qwen3.6-35B-A3BQwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 188,8 tok/s |
3.560 | 188,8 | 6336 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 46 | Qwen-AgentWorld-35B-A3BQwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 187,2 tok/s |
3.596 | 187,2 | 6276 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
| 47 | Qwen3-VL-30B-A3B-InstructQwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 186,9 tok/s |
4.641 | 186,9 | 2689 ms | 5× | llama.cppgodclawQ4_K_M | Details → |
| 48 | gpt-oss-20bOpenAI | NVIDIA GeForce RTX 3090 Ti | 32x AMD Ryzen 9 8945HX with Radeon Graphics 92 GB | 176,2 tok/s |
8.074 | 176,2 | 245 ms | 1× | vLLMwebsocket | Details → |
| 49 | MiniMax-M2.7-AWQ-4bitMiniMax | NVIDIA GB10 (DGX Spark) | NVIDIA Grace 120 GB | 174,2 tok/s |
5.042 | 174,2 | 47619 ms | 10× | vLLMgodclawAWQ | Details → |
| 50 | Qwen3-30B-A3B-Instruct-2507Qwen (Alibaba) | 3x AMD Radeon AI PRO R9700 | 32x AMD Ryzen Threadripper PRO 7955WX 16-Cores 184 GB | 173,8 tok/s |
6.737 | 173,8 | 4318 ms | 10× | llama.cppgodclawQ4_K_M | Details → |
Ranking aus veroeffentlichten Benchmarks. Sortiert nach Generation-Durchsatz. Wie wird gemessen?