🏆
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 | minimax-m2-awq230BMiniMax Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 547,88 tok/s TG Prefill 4.819 · TTFT 20.318 ms | 10× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawUD-Q4_K_M | Details → | |
| 2 | minimax-m2-awq230BMiniMax Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 310,36 tok/s TG Prefill 4.459 · TTFT 6.645 ms | 5× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawUD-Q4_K_M | Details → | |
| 3 | minimax-m2-awq230BMiniMax Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 112,64 tok/s TG Prefill 2.405 · TTFT 972 ms | 1× | 3x NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation EditionAMD Ryzen Threadripper PRO 9965WX 24-Cores · 125 GB RAM | llama.cppgodclawUD-Q4_K_M | Details → |