🏆
Ranglisten
LLM Leaderboard
Reale Messergebnisse aller getesteten Sprachmodelle auf echter, benannter Hardware – bewertet nach roher Geschwindigkeit (Performance in Token pro Sekunde, Prefill und Zeit bis zum ersten Token) und nach praktischer Aufgaben-Qualität in vollständigen Agenten- und Chat-Abläufen (Harness). Wähle unten einen Benchmark-Typ oder filtere nach Modell, Hersteller und Hardware, um genau zu sehen, was jeweils geprüft wird und wie die Ergebnisse zustande kommen.
⚡ Performance (tok/s)🤖 Harness-Qualität👥 Concurrency🖥️ echte Hardware
| # | Modell / Hersteller | Messwerte | Parallel | GPU / CPU / RAM | Runtime | ||
|---|---|---|---|---|---|---|---|
| 1 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 830,46 tok/s TG Prefill 3.771 · TTFT 14.928 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | hermes_cli | Details → | |
| 2 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 829,45 tok/s TG Prefill 3.778 · TTFT 14.891 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | hermes_cli | Details → | |
| 3 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 828,36 tok/s TG Prefill 3.763 · TTFT 14.940 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacppQ4_K_M | Details → | |
| 4 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 572,10 tok/s TG Prefill 10.901 · TTFT 1.915 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllm mxfp4 | Details → | |
| 5 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 467,09 tok/s TG Prefill 3.349 · TTFT 5.143 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacppQ4_K_M | Details → | |
| 6 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 451,05 tok/s TG Prefill 32.005 · TTFT 864 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllm | Details → | |
| 7 | Nemotron-3-Nano-30B-A3B30BNVIDIA PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 404,72 tok/s TG Prefill 4.793 · TTFT 5.226 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacppIQ4_NL | Details → | |
| 8 | Gemma-4-26B-A4B-it26BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 397,18 tok/s TG Prefill 3.476 · TTFT 5.916 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → | |
| 9 | gemma-4-12B-it12BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 372,40 tok/s TG Prefill 1.508 · TTFT 13.832 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMllamacppAWQ | Details → | |
| 10 | gemma-4-E4B-it8BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 346,29 tok/s TG Prefill 16.028 · TTFT 1.776 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllm bf16 | Details → | |
| 11 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 308,36 tok/s TG Prefill 9.336 · TTFT 1.062 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllm mxfp4 | Details → | |
| 12 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 306,80 tok/s TG Prefill 9.358 · TTFT 1.060 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllm mxfp4 | Details → | |
| 13 | Nemotron-3-Nano-30B-A3B30BNVIDIA PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 300,10 tok/s TG Prefill 3.941 · TTFT 2.860 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacppIQ4_NL | Details → | |
| 14 | Gemma-4-26B-A4B-it26BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 287,13 tok/s TG Prefill 3.082 · TTFT 3.211 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → | |
| 15 | Codestral-22B-v0.122BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 277,86 tok/s TG Prefill 3.781 · TTFT 10.183 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllmAWQ | Details → | |
| 16 | Devstral-Small-2-24B-Instruct-251224BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 247,69 tok/s TG Prefill 5.149 · TTFT 6.214 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → | |
| 17 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 228,02 tok/s TG Prefill 11.489 · TTFT 2.605 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllm | Details → | |
| 18 | Codestral-22B-v0.122BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 214,10 tok/s TG Prefill 4.633 · TTFT 8.462 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → | |
| 19 | Nemotron-3-Nano-30B-A3B30BNVIDIA PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 196,02 tok/s TG Prefill 3.211 · TTFT 670 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacppIQ4_NL | Details → | |
| 20 | Codestral-22B-v0.122BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 184,33 tok/s TG Prefill 2.857 · TTFT 5.676 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllmAWQ | Details → | |
| 21 | gemma-4-12B-it12BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 183,69 tok/s TG Prefill 2.125 · TTFT 4.654 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMllamacppAWQ | Details → | |
| 22 | gemma-4-E4B-it8BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 177,40 tok/s TG Prefill 11.718 · TTFT 941 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllm bf16 | Details → | |
| 23 | gemma-4-31B-it31BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 174,07 tok/s TG Prefill 1.206 · TTFT 16.925 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → | |
| 24 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 162,37 tok/s TG Prefill 2.928 · TTFT 659 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | hermes_cli | Details → | |
| 25 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 162,17 tok/s TG Prefill 2.928 · TTFT 660 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacppQ4_K_M | Details → | |
| 26 | gemma-4-31B-it31BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 156,49 tok/s TG Prefill 182 · TTFT 97.092 ms | 10× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llama.cpp q3_k_m, q4-kv | Details → | |
| 27 | Gemma-4-26B-A4B-it26BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 125,87 tok/s TG Prefill 3.206 · TTFT 618 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → | |
| 28 | Codestral-22B-v0.122BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 125,21 tok/s TG Prefill 3.410 · TTFT 4.809 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → | |
| 29 | Devstral-Small-2-24B-Instruct-251224BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 123,94 tok/s TG Prefill 3.957 · TTFT 3.439 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → | |
| 30 | gemma-4-12B-it12BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 83,56 tok/s TG Prefill 2.303 · TTFT 859 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMllamacppAWQ | Details → | |
| 31 | gemma-4-31B-it31BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 81,35 tok/s TG Prefill 998 · TTFT 9.924 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → | |
| 32 | gemma-4-31B-it31BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 74,79 tok/s TG Prefill 1.062 · TTFT 9.322 ms | 5× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llama.cpp q3_k_m, q4-kv | Details → | |
| 33 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 61,64 tok/s TG Prefill 5.671 · TTFT 366 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllm mxfp4 | Details → | |
| 34 | Codestral-22B-v0.122BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 57,55 tok/s TG Prefill 1.941 · TTFT 1.402 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → | |
| 35 | Devstral-Small-250724BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 57,12 tok/s TG Prefill 3.392 · TTFT 922 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → | |
| 36 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 47,58 tok/s TG Prefill 10.520 · TTFT 189 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllmBF16 | Details → | |
| 37 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 46,20 tok/s TG Prefill 11.058 · TTFT 180 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllm | Details → | |
| 38 | gemma-4-E2B-it5BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 46,01 tok/s TG Prefill 11.757 · TTFT 169 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllmBF16 | Details → | |
| 39 | Codestral-22B-v0.122BMistral AI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 45,38 tok/s TG Prefill 1.437 · TTFT 1.875 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vllm | Details → | |
| 40 | gemma-4-E4B-it8BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 37,86 tok/s TG Prefill 6.286 · TTFT 315 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | vLLMvllm bf16 | Details → | |
| 41 | gemma-4-31B-it31BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 30,10 tok/s TG Prefill 1.020 · TTFT 1.964 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llama.cpp q3_k_m, q4-kv | Details → | |
| 42 | gemma-4-31B-it31BGoogle PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 28,74 tok/s TG Prefill 938 · TTFT 2.110 ms | 1× | NVIDIA GeForce RTX 3090 TiAMD Ryzen 5 5600X 6-Core Processor · 30 GB RAM | llamacpp | Details → |