benchmarks

Browser-based benchmark for testing LLM inference endpoints - speed, latency, accuracy, and cost.