⚡
Performance-Benchmark
Wie schnell ist das Modell?
Roher Durchsatz auf echter Hardware – wie viele Token ein Modell pro Sekunde erzeugt, wie schnell es den Prompt verarbeitet und wie kurz die Zeit bis zum ersten Token ausfällt. Je höher, desto besser.
⚡ Generation (tok/s)🔄 Prefill⏱️ TTFT👥 Concurrency
| # | Modell / Hersteller | Messwerte | Parallel | GPU / CPU / RAM | Runtime | ||
|---|---|---|---|---|---|---|---|
| 1 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 200,47 tok/s TG Prefill 5.303 · TTFT 7.750 ms | 10× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliMXFP4 | Details → | |
| 2 | Nemotron-3.5-Lightning-30B-A3B30BNVIDIA PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 146,86 tok/s TG Prefill 2.806 · TTFT 12.191 ms | 10× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_0 | Details → | |
| 3 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 141,98 tok/s TG Prefill 4.212 · TTFT 12.237 ms | 10× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_K_M | Details → | |
| 4 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 124,89 tok/s TG Prefill 2.476 · TTFT 6.821 ms | 5× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_K_M | Details → | |
| 5 | Nemotron-3.5-Lightning-30B-A3B30BNVIDIA PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 121,80 tok/s TG Prefill 1.943 · TTFT 7.408 ms | 5× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_0 | Details → | |
| 6 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 119,73 tok/s TG Prefill 3.482 · TTFT 4.698 ms | 5× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliMXFP4 | Details → | |
| 7 | gpt-oss-120b120BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 111,85 tok/s TG Prefill 2.041 · TTFT 17.689 ms | 10× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliMXFP4 | Details → | |
| 8 | Qwen3-30B-A3B-Thinking-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 106,99 tok/s TG Prefill 5.833 · TTFT 17.334 ms | 10× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_K_M | Details → | |
| 9 | gpt-oss-120b120BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 79,06 tok/s TG Prefill 1.478 · TTFT 11.284 ms | 5× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliMXFP4 | Details → | |
| 10 | Qwen3-30B-A3B-Thinking-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 75,08 tok/s TG Prefill 1.575 · TTFT 1.551 ms | 1× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_K_M | Details → | |
| 11 | Nemotron-3.5-Lightning-30B-A3B30BNVIDIA PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 72,08 tok/s TG Prefill 1.293 · TTFT 1.669 ms | 1× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_0 | Details → | |
| 12 | Qwen3-30B-A3B-Thinking-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 71,97 tok/s TG Prefill 3.029 · TTFT 9.456 ms | 5× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_K_M | Details → | |
| 13 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 70,74 tok/s TG Prefill 2.529 · TTFT 846 ms | 1× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliMXFP4 | Details → | |
| 14 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 67,87 tok/s TG Prefill 1.288 · TTFT 1.499 ms | 1× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_K_M | Details → | |
| 15 | Nemotron-3.5-Lightning-30B-A3B30BNVIDIA PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 66,20 tok/s TG Prefill 1.300 · TTFT 1.660 ms | 1× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_0 | Details → | |
| 16 | Qwen3.6-35B-A3B35BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 64,50 tok/s TG Prefill 1.366 · TTFT 1.420 ms | 1× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_K_M | Details → | |
| 17 | gpt-oss-20b20BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 64,41 tok/s TG Prefill 2.523 · TTFT 858 ms | 1× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliMXFP4 | Details → | |
| 18 | Qwen3-30B-A3B-Thinking-250730BQwen (Alibaba) PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 63,15 tok/s TG Prefill 1.472 · TTFT 1.659 ms | 1× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliQ4_K_M | Details → | |
| 19 | gpt-oss-120b120BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 51,11 tok/s TG Prefill 899 · TTFT 2.279 ms | 1× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliMXFP4 | Details → | |
| 20 | gpt-oss-120b120BOpenAI PerformancebenchmarkTimebench 3 - Kombi (Prefill + Generation) | 48,95 tok/s TG Prefill 919 · TTFT 2.243 ms | 1× | NVIDIA RTX SPARK (N1X)NVIDIA RTX Spark N1X (6144-core GPU, 20-core CPU) · 128 GB RAM | llama.cppopenclaw_cliMXFP4 | Details → |