Loading the SOTA2 catalog…
LLM Inference Serving Performance on 1024 Input / 32 Output Context benchmark leaderboard · SOTA2 Research