OpenRouter Adds Tencent Hy4 Preview

What shipped? Tencent's Hy4 preview model is now available on OpenRouter — a 770B-parameter Mixture-of-Experts model with 49B active parameters per token.

What changed? Hy4 preview supports up to 1,048,576 tokens of context and 64,000 completion tokens. Pricing is competitive: $0.83/M tokens input, $2.50/M tokens output, with cached reads at $0.042/M. The model is designed for coding agents, complex tool-use workflows, and productivity tasks. Free tier via Ling 3.0 Flash Fin also landed yesterday on OpenRouter as inclusionai/ling-3.0-flash-fin:free — a finance-focused 124B MoE model at zero cost.

Why a builder cares: Hy4 enters the large-context MoE arena at a price point that undercuts equivalently-sized rivals. For agentic workflows that need million-token context windows without breaking the budget, this is a new option worth benchmarking against Qwen and DeepSeek.