🏆
Rankings
LLM Leaderboard
Real measurements of every tested language model on real, named hardware – rated by raw speed (performance in tokens per second, prefill and time to first token) and by practical task quality in complete agent and chat runs (harness). Pick a benchmark type below or filter by model, maker and hardware to see exactly what is tested and how the results are produced.
⚡ Performance (tok/s)🤖 Harness quality👥 Concurrency🖥️ real hardware
| # | Model / Maker | Metrics | Parallel | GPU / CPU / RAM | Runtime | ||
|---|---|---|---|---|---|---|---|
| 1 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 1.608,46 tok/s TG Prefill 31.452 · TTFT 1.633 ms | 10× | NVIDIA GeForce RTX 5070 TiAMD Ryzen Threadripper PRO 5975WX 32-Cores · 247 GB RAM | vLLMgodclaw | Details → | |
| 2 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 973,54 tok/s TG Prefill 23.599 · TTFT 460 ms | 5× | NVIDIA GeForce RTX 5070 TiAMD Ryzen Threadripper PRO 5975WX 32-Cores · 247 GB RAM | vLLMgodclaw | Details → | |
| 3 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 228,18 tok/s TG Prefill 11.075 · TTFT 179 ms | 1× | NVIDIA GeForce RTX 5070 TiAMD Ryzen Threadripper PRO 5975WX 32-Cores · 247 GB RAM | vLLMgodclaw | Details → | |
| 4 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 90,00 tok/s TG Prefill 3.340 · TTFT 6.061 ms | 10× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 5 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 89,68 tok/s TG Prefill 3.419 · TTFT 5.899 ms | 10× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 6 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 89,50 tok/s TG Prefill 3.479 · TTFT 5.849 ms | 10× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 7 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 88,97 tok/s TG Prefill 3.386 · TTFT 5.965 ms | 10× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 8 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 79,49 tok/s TG Prefill 758 · TTFT 13.043 ms | 5× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawUD-Q4_K_M | Details → | |
| 9 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 78,82 tok/s TG Prefill 767 · TTFT 12.893 ms | 5× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawUD-Q4_K_M | Details → | |
| 10 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 78,62 tok/s TG Prefill 760 · TTFT 13.000 ms | 5× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 11 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 78,25 tok/s TG Prefill 756 · TTFT 13.064 ms | 5× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 12 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 77,81 tok/s TG Prefill 881 · TTFT 23.158 ms | 10× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawUD-Q4_K_M | Details → | |
| 13 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 76,93 tok/s TG Prefill 887 · TTFT 22.982 ms | 10× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawUD-Q4_K_M | Details → | |
| 14 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 76,61 tok/s TG Prefill 865 · TTFT 23.569 ms | 10× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 15 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 73,70 tok/s TG Prefill 874 · TTFT 23.415 ms | 10× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 16 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 71,78 tok/s TG Prefill 3.102 · TTFT 7.392 ms | 10× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → | |
| 17 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 64,64 tok/s TG Prefill 3.220 · TTFT 6.301 ms | 10× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → | |
| 18 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 59,15 tok/s TG Prefill 2.977 · TTFT 3.328 ms | 5× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 19 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 59,06 tok/s TG Prefill 2.935 · TTFT 3.376 ms | 5× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 20 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 55,40 tok/s TG Prefill 3.214 · TTFT 6.335 ms | 10× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → | |
| 21 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 54,10 tok/s TG Prefill 3.151 · TTFT 6.474 ms | 10× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → | |
| 22 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 45,65 tok/s TG Prefill 2.758 · TTFT 3.599 ms | 5× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → | |
| 23 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 45,52 tok/s TG Prefill 2.714 · TTFT 3.657 ms | 5× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 24 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 33,49 tok/s TG Prefill 2.798 · TTFT 3.540 ms | 5× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → | |
| 25 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 32,31 tok/s TG Prefill 782 · TTFT 2.530 ms | 1× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 26 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 32,11 tok/s TG Prefill 726 · TTFT 2.722 ms | 1× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawUD-Q4_K_M | Details → | |
| 27 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 31,98 tok/s TG Prefill 776 · TTFT 2.550 ms | 1× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 28 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 31,75 tok/s TG Prefill 736 · TTFT 2.686 ms | 1× | AMD Radeon 8060S GraphicsAMD RYZEN AI MAX+ 395 w/ Radeon 8060S · 31 GB RAM | llama.cppgodclawUD-Q4_K_M | Details → | |
| 29 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 28,96 tok/s TG Prefill 2.290 · TTFT 4.418 ms | 5× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → | |
| 30 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 27,21 tok/s TG Prefill 2.637 · TTFT 3.761 ms | 5× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → | |
| 31 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 26,45 tok/s TG Prefill 2.305 · TTFT 4.311 ms | 5× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 32 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 12,34 tok/s TG Prefill 1.975 · TTFT 1.000 ms | 1× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 33 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 12,34 tok/s TG Prefill 1.950 · TTFT 1.013 ms | 1× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 34 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 12,34 tok/s TG Prefill 1.963 · TTFT 1.006 ms | 1× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → | |
| 35 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 12,33 tok/s TG Prefill 1.967 · TTFT 1.004 ms | 1× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 36 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 12,33 tok/s TG Prefill 1.951 · TTFT 1.012 ms | 1× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_0 | Details → | |
| 37 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 12,33 tok/s TG Prefill 1.949 · TTFT 1.014 ms | 1× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → | |
| 38 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 12,32 tok/s TG Prefill 1.969 · TTFT 1.002 ms | 1× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → | |
| 39 | gemma-4-E2B-itGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 5,76 tok/s TG Prefill 843 · TTFT 2.406 ms | 1× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawQ4_K_M | Details → |