Tencent released Hy4 Preview, a 770B-A49B parameter model offering performance comparable to leading Chinese AI models (DeepSeek V4 Pro, Kimi K3, GLM5.3, Qwen3.8 Max) while maintaining aggressive pricing at $0.834–$2.501 per million tokens—with exceptionally low cache rates ($0.042/M tokens). This makes Hy4 cost-competitive for agentic AI workloads, particularly relevant for SMEs exploring large-scale LLM deployment without prohibitive inference costs.