🏆
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 | Meta-Llama-3.1-8B-Instruct8BMeta Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 317,00 tok/s TG Prefill 14.849 · TTFT 2.492 ms | 10× | 2x NVIDIA RTX A6000 iAMD EPYC 7203P 8-Core Processor · 126 GB RAM | vLLMgodclaw | Details → | |
| 2 | Meta-Llama-3.1-8B-Instruct8BMeta Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 183,35 tok/s TG Prefill 11.565 · TTFT 1.219 ms | 5× | 2x NVIDIA RTX A6000 iAMD EPYC 7203P 8-Core Processor · 126 GB RAM | vLLMgodclaw | Details → | |
| 3 | Meta-Llama-3.1-8B-Instruct8BMeta Performance benchmarkTimebench 3 - Kombi (Prefill + Generation) | 42,39 tok/s TG Prefill 5.909 · TTFT 412 ms | 1× | 2x NVIDIA RTX A6000 iAMD EPYC 7203P 8-Core Processor · 126 GB RAM | vLLMgodclaw | Details → |

