Ranglisten
LLM Leaderboard
Filtere reale Messergebnisse nach Benchmarktyp und Systemkonfiguration.
| # | Modell / Hersteller | Typ | Messwerte | GPU / CPU / RAM | Runtime | Datum | ||
|---|---|---|---|---|---|---|---|---|
| 1 | MiniMax-M2.5MiniMaxTool Usage Standard 1.0 | Harnessbenchmark | 81,4 % 2.427/2.980 Pkt · 16/16 Aufg. | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition34x AMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | vLLMgodclaw BF16 | 13.07.2026 | Details → | |
| 2 | MiniMax-M2.5MiniMaxTool Usage Standard 1.0 | Harnessbenchmark | 78,5 % 2.340/2.980 Pkt · 16/16 Aufg. | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition34x AMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | openclaw_cli | 06.04.2026 | Details → | |
| 3 | MiniMax-M2.5-BF16-INT4-AWQMiniMaxTool Usage V 1.0 | Harnessbenchmark | 70,7 % 1.245/1.760 Pkt · 15/15 Aufg. | NVIDIA GB10NVIDIA Grace · 120 GB RAM | openclaw_cli | 04.04.2026 | Details → | |
| 4 | MiniMax-M2.5MiniMaxTool Usage Standard 1.0 | Harnessbenchmark | 68,2 % 2.033/2.980 Pkt · 16/16 Aufg. | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition34x AMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | vLLMgodclaw BF16 | 13.07.2026 | Details → | |
| 5 | MiniMax-M2.5MiniMaxTools Usage Mini 1.0 | Harnessbenchmark | 65,0 % 130/200 Pkt · 2/2 Aufg. | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition34x AMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | godclaw | 08.07.2026 | Details → | |
| 6 | MiniMax-M2.5MiniMaxPerformancetest Small 1.0 | Performancebenchmark | 62,70 tok/s Prefill 368 · TTFT 144 ms | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition34x AMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | vLLMgodclaw BF16 | 20.07.2026 | Details → | |
| 7 | MiniMax-M2.5MiniMaxTool Usage V 1.0 | Harnessbenchmark | 61,3 % 1.079/1.760 Pkt · 15/15 Aufg. | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition34x AMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | openclaw_cli | 04.04.2026 | Details → | |
| 8 | MiniMax-M2.7-AWQ-4bitMiniMaxTool-Parcours Dossier 40 V1.0 | Harnessbenchmark | 61,1 % 3.641/5.960 Pkt · 40/40 Aufg. | NVIDIA GB10NVIDIA Grace · 120 GB RAM | vLLMgodclaw AWQ | 19.07.2026 | Details → | |
| 9 | MiniMax-M2.5MiniMaxTool-Parcours Dossier 40 V1.0 | Harnessbenchmark | 60,8 % 3.622/5.960 Pkt · 34/40 Aufg. | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition34x AMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | vLLMgodclaw BF16 | 19.07.2026 | Details → | |
| 10 | MiniMax-M2.5-BF16-INT4-AWQMiniMaxTool Usage Standard 1.0 | Harnessbenchmark | 59,1 % 1.762/2.980 Pkt · 16/16 Aufg. | NVIDIA GB10NVIDIA Grace · 120 GB RAM | openclaw_cli | 06.04.2026 | Details → | |
| 11 | MiniMax-M2.5MiniMaxTool Usage Standard 1.0 | Harnessbenchmark | 57,5 % 1.714/2.980 Pkt · 16/16 Aufg. | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition34x AMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | vLLMgodclaw BF16 | 15.07.2026 | Details → | |
| 12 | MiniMax-M2.5MiniMaxTools Usage Mini 1.0 | Harnessbenchmark | 53,0 % 106/200 Pkt · 2/2 Aufg. | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition34x AMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | openclaw_cli | 03.04.2026 | Details → | |
| 13 | MiniMax-M2.5-BF16-INT4-AWQMiniMaxTextgenerierung V 1.0 | Harnessbenchmark | 52,0 % 790/1.520 Pkt · 15/15 Aufg. | NVIDIA GB10NVIDIA Grace · 120 GB RAM | openclaw_cli | 04.04.2026 | Details → | |
| 14 | MiniMax-M2.7-AWQ-4bitMiniMaxPerformancetest Small 1.0 | Performancebenchmark | 34,37 tok/s Prefill 503 · TTFT 105 ms | NVIDIA GB10NVIDIA Grace · 120 GB RAM | vLLMgodclaw AWQ | 20.07.2026 | Details → | |
| 15 | MiniMax-M2.7-AWQ-4bitMiniMaxTool Usage Standard 1.0 | Harnessbenchmark | 28,8 % 858/2.980 Pkt · 16/16 Aufg. | NVIDIA GB10NVIDIA Grace · 120 GB RAM | godclaw | 15.07.2026 | Details → | |
| 16 | MiniMax-M2.5MiniMaxTools Usage Mini 1.0 | Harnessbenchmark | 10,0 % 10/100 Pkt · 1/1 Aufg. | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition34x AMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | openclaw_cli | 03.04.2026 | Details → | |
| 17 | Minimax-2.5MiniMaxTimebench 3 - Kombi (Prefill + Generation) | Performancebenchmark | 6,46 tok/s Prefill 119 · TTFT 21.346 ms | 3x AMD Radeon AI PRO R970032x AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | llama.cppgodclaw | 22.07.2026 | Details → | |
| 18 | MiniMax-M2.7MiniMaxTimebench 3 - Kombi (Prefill + Generation) | Performancebenchmark | 3,00 tok/s Prefill 139 · TTFT 15.493 ms | 3x AMD Radeon AI PRO R970032x AMD Ryzen Threadripper PRO 7955WX 16-Cores · 184 GB RAM | llama.cppgodclaw Q4_K_M | 22.07.2026 | Details → | |
| 19 | MiniMax-M2.5MiniMaxTimebench 3 - Kombi (Prefill + Generation) | Performancebenchmark | 0,84 tok/s Prefill 22 · TTFT 98.716 ms | Keine GPU (CPU-only)51x Intel(R) Xeon(R) CPU E5-4620 v2 @ 2.60GHz · 504 GB RAM | llama.cppgodclaw Q4_K_M | 22.07.2026 | Details → |