Ranglisten
LLM Leaderboard
Filtere reale Messergebnisse nach Benchmarktyp und Systemkonfiguration.
| # | Modell / Hersteller | Typ | Messwerte | GPU / CPU / RAM | Runtime | Datum | ||
|---|---|---|---|---|---|---|---|---|
| 1 | NVIDIA-Nemotron-3-Nano-4BNVIDIATimebench 3 - Kombi (Prefill + Generation) | Performancebenchmark | 4,94 tok/s Prefill 41 · TTFT 62.079 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 → | |
| 2 | Qwen3.6-35B-A3BQwen (Alibaba)Timebench 3 - Kombi (Prefill + Generation) | Performancebenchmark | 4,29 tok/s Prefill 62 · TTFT 31.435 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 → | |
| 3 | Qwen2.5-72B-Instruct-Q4_K_M.ggufQwen (Alibaba)Performancetest Small 1.0 | Performancebenchmark | 2,28 tok/s Prefill 81 · TTFT 569 ms | Keine GPU (CPU-only)51x Intel(R) Xeon(R) CPU E5-4620 v2 @ 2.60GHz · 504 GB RAM | godclaw | 20.07.2026 | Details → | |
| 4 | 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 → |