Global Leaderboard

High-performance inference tracking across multiple hardware architectures. Results are aggregated from verified community submissions.

RANK	DEVICE NAME	SYSTEM INFO	BACKEND	TYPE	SCORE (TOK/S)
#1	NVIDIA H100 PCIe 80GB HBM3	Ubuntu 22.04 2TB RAM	TensorRT-LLM	developer_board	4,820.5
#2	NVIDIA RTX 4090 24GB GDDR6X	Windows 11 128GB RAM	vLLM	video_camera_front	1,245.8
#3	Apple M3 Max 40-core GPU	macOS 14.2 128GB Unified	MLX	memory	890.2
#4	AMD Radeon RX 7900 XTX 24GB GDDR6	Ubuntu 22.04 64GB RAM	ROCm	video_camera_front	812.5
#5	AMD EPYC 9654 96 Cores / 192 Threads	Debian 12 768GB RAM	llama.cpp	developer_board	450.1

Showing 1-5 of 1,204 results