The mysterious Hy3 LLM is topping OpenRouter Model Rankings by a large margin

2026-05-26 · Source: Max Woolf's Blog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Software Development & Engineering · Depth: Intermediate, short

Summary

The OpenRouter AI Model Rankings, retrieved May 25, 2026, show Tencent's Hy3 preview LLM unexpectedly topping token usage by over 50% compared to models like Claude, despite its seemingly inferior quality and higher effective cost. Hy3 preview, priced at \$0.066/1M input tokens from SiliconFlow, has a 44% cache read cost, resulting in an effective price of \$0.034/1M input tokens. In contrast, DeepSeek V4 Flash, when served directly by DeepSeek, offers a 2% cache read cost, leading to a significantly lower effective price of \$0.018/1M input tokens. The article highlights that 98% of LLM API costs are now input tokens, making prompt caching crucial and stated prices misleading. The author remains puzzled by Hy3's popularity, suggesting a single large, non-agentic app might be its primary user.

Key takeaway

For AI Engineers optimizing LLM costs, you must look beyond stated API prices. Effective pricing, heavily influenced by prompt caching and provider-specific cache read costs, can nearly double or halve your actual spend. Prioritize providers like DeepSeek directly for DeepSeek V4 Flash, which offers a 2% cache read cost, significantly undercutting models like Hy3 preview. Always consult OpenRouter's effective pricing table and consider data policy implications before deployment.

Key insights

Effective LLM pricing, driven by prompt caching and high input token usage, significantly deviates from stated costs, influencing model adoption.

Principles

• LLM API costs are 98% input tokens.
• Prompt caching significantly alters effective pricing.
• Cache read costs vary widely by provider.

In practice

• Consult OpenRouter's effective pricing table.
• Evaluate direct API keys for better caching.
• Review provider data policy on prompt training.

Topics

• LLM Economics
• OpenRouter Rankings
• Prompt Caching
• DeepSeek V4 Flash
• Hy3 Preview
• API Pricing

Best for: MLOps Engineer, CTO, VP of Engineering/Data, AI Engineer, Machine Learning Engineer, Director of AI/ML

Related on AIssential

• DeepSeek V4: One Million Tokens, Three Thinking Modes, and the First Real Hands-On Reports — DeepSeek-V4 optimizes long-context LLMs via novel attention and training stability, offering tiered reasoning and cost-effective variants.
• Five things we learned trimming LibreChat’s LLM bill — Strategic LLM cost optimization requires understanding provider pricing, intelligent model routing, and careful resource management.
• Gemini 3.5 Flash: more expensive, but Google plan to use it for everything — Gemini 3.5 Flash signals a new phase of higher LLM pricing and broad internal deployment across Google's ecosystem.
• Accelerating LLM Inference with Prompt Caching for Open‑Source Models on Databricks — Prompt caching for LLMs significantly reduces inference latency and increases throughput by reusing identical prompt prefixes.
• DeepSeek just confirmed that their 75% promo discount for the V4-Pro API is actually becoming the permanent price — DeepSeek permanently reduced V4-Pro API prices, significantly undercutting competitors and intensifying the LLM market's price war.
• How an Unexpected Reddit Spike Forced Me to Learn Prompt Caching the Hard Way — Multi-agent LLM pipelines offer significant prompt caching benefits, especially when context is passed sequentially between agents.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Max Woolf's Blog.