Created bymario-alka.dePowered bynoob2claw.detricoma.de
🏆
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
Reset
Metric:
#Model / MakerMetricsParallelGPU / CPU / RAMRuntime
1MiniMax-M2.5230BMiniMax Performance benchmarkTimebench 3 - Kombi (Prefill + Generation)
103,49 tok/s TG
Prefill 7.016 · TTFT 5.294 ms
10×NVIDIA DGX SPARK (GB10)NVIDIA Grace · 120 GB RAMvLLMgodclawAWQ
2MiniMax-M2.5230BMiniMax Performance benchmarkTimebench 3 - Kombi (Prefill + Generation)
77,74 tok/s TG
Prefill 5.460 · TTFT 2.641 ms
5×NVIDIA DGX SPARK (GB10)NVIDIA Grace · 120 GB RAMvLLMgodclawAWQ
3MiniMax-M2.5230BMiniMax Performance benchmarkTimebench 3 - Kombi (Prefill + Generation)
32,23 tok/s TG
Prefill 2.184 · TTFT 1.050 ms
1×NVIDIA DGX SPARK (GB10)NVIDIA Grace · 120 GB RAMvLLMgodclawAWQ
4MiniMax-M2.5230BMiniMax Harness benchmarkTool Usage Standard 1.0
1.762 Pkt
59,1% · 16/16 Aufg. · 526,0 s
1×NVIDIA DGX SPARK (GB10)NVIDIA Grace · 120 GB RAMopenclaw_cliBF16-INT4-AWQ
5MiniMax-M2.5230BMiniMax Harness benchmarkTextgenerierung V 1.0
790 Pkt
52,0% · 15/15 Aufg. · 261,0 s
1×NVIDIA DGX SPARK (GB10)NVIDIA Grace · 120 GB RAMopenclaw_cliBF16-INT4-AWQ
6MiniMax-M2.5230BMiniMax Harness benchmarkTool Usage V 1.0
1.245 Pkt
70,7% · 15/15 Aufg. · 301,0 s
1×NVIDIA DGX SPARK (GB10)NVIDIA Grace · 120 GB RAMopenclaw_cliBF16-INT4-AWQ