🏆
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-4-12B-it12BQ4_K_MGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 89,80 tok/s TG Prefill 877 · TTFT 23.528 ms | 10× | 2x NVIDIA GeForce RTX 2060Intel(R) Core(TM) i5-7400 CPU @ 3.00GHz · 7 GB RAM | llama.cppopenclaw_cli | Details → | |
| 2 | gemma-4-12B-it12BQ4_K_MGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 60,61 tok/s TG Prefill 776 · TTFT 12.750 ms | 5× | 2x NVIDIA GeForce RTX 2060Intel(R) Core(TM) i5-7400 CPU @ 3.00GHz · 7 GB RAM | llama.cppopenclaw_cli | Details → | |
| 3 | gemma-4-12B-it12BQ4_K_MGoogle Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 30,34 tok/s TG Prefill 684 · TTFT 2.896 ms | 1× | 2x NVIDIA GeForce RTX 2060Intel(R) Core(TM) i5-7400 CPU @ 3.00GHz · 7 GB RAM | llama.cppopenclaw_cli | Details → |