Tencent launches Hy4 flash, a 224B MoE model at $0.07 per million input tokens
Tencent released Hy4 flash on November 16, 2026, a 224-billion-parameter mixture-of-experts model with 12 billion active parameters, multimodal inputs and context of up to 1 million tokens, priced at about $0.070 per mil
Tencent released Hy4 flash on November 16, 2026, a 224-billion-parameter mixture-of-experts model with 12 billion active parameters, multimodal inputs and context of up to 1 million tokens, priced at about $0.070 per million input tokens and $0.281 per million output tokens on Tencent Cloud TokenHub.
What changed with Hy4 flash
Tencent released Hy4 flash on November 16, 2026, as the most cost-effective model in the Hy family. It has 224 billion total parameters, including 64 billion Engram parameters, with only 12 billion active per token, multimodal inputs and context of up to 1 million tokens.
Hy4 flash is available now across Tencent Cloud TokenHub, WorkBuddy, Yuanbao, WeCom, QQ, ima and Marvis, with OpenRouter integration coming soon. Tencent also plans to open-source Hy4 flash, with no date given; as of November 16, 2026, Artificial Analysis had not published a page for the model.
Pricing is about $0.070 per million input tokens, $0.281 per million output tokens and $0.003 per million cached input tokens, with decoding speeds exceeding 150 tokens per second using 4-bit quantization and inference optimizations. On WorkIndex, Tencent's internal benchmark drawn from WorkBuddy-style office tasks, Hy4 flash delivers a per-task cost of just 1/20 that of full Hy4. Through November 30, 2026, Hy4 flash is free in WorkBuddy and 50% off standard pricing via TokenHub.
Updates
- November 16, 2026: Hy4 Flash is also available on OpenRouter as tencent/hy4-flash, served by Tencent Cloud at $0.035 per million input tokens, $0.1405 per million output tokens and $0.0015 per million cache reads, with a 50% introductory discount and the same 1-million-token context. It is also available on Vercel AI Gateway as tencent/hy4-flash, with an introductory 50% discount on input, cached and output tokens through November 29, 2026, and setup for coding agents including Claude Code, Codex and Cursor.
Hy4 flash by the numbers
- Input token price ($ per million tokens): $0.07 (Nov 16, 2026)
- Output token price ($ per million tokens): $0.281 (Nov 16, 2026)
- Context window: 1,000,000 tokens (Nov 16, 2026)
- Decoding speed: More than 150 tokens/s (Nov 16, 2026)
Why Hy4 flash matters for business
Businesses can route everyday office work β deep research, document processing, spreadsheet repair, slide-deck generation and simple computer use β to the Hy4 flash API at a fraction of flagship prices, reserving full Hy4 or other frontier models for long, complex jobs. The free WorkBuddy window and half-price TokenHub access through November 30, 2026, make it cheap to test on real workloads before committing.
Sources
Originally published by Dev.to AI. Aggregated on AIWithGhost for educational purposes β full credit and traffic to the original publisher.