🇨🇳China·AI News·19 Nov 2026·via Tech Node

    Tencent says Hunyuan Hy4 flash surpasses 50 trillion tokens in first-day usage

    Tencent's Hunyuan Hy4 flash AI model processed over 50 trillion tokens on its first day after launching on November 16, according to the company. This marked the highest first-day usage among models in the Hunyuan series, with Tencent reporting its strong debut also led to a top ranking on OpenRouter’s daily LLM leaderboard on November 17.

    Tencent’s Hunyuan Hy4 flash AI model, launched on November 16, processed more than 50 trillion tokens on its first day. This figure, reported by Tencent, represents the highest initial usage among models in the Hunyuan series. The company also stated that on November 17, Hy4 flash ranked first on OpenRouter’s daily LLM leaderboard. This strong debut follows an earlier surge in demand for its predecessor, Hy4, which led to users hitting usage limits when it debuted on Tencent’s WorkBuddy platform on November 4. At that time, Tencent expanded its inference clusters tenfold and increased free usage quotas. For Hy4 flash, the team appears to have made more extensive preparations for high concurrency, as no widespread usage-limit issues have been reported.

    The Hy4 flash model, which Tencent states uses a hybrid activation architecture, has a total of 224 billion parameters, including 64 billion Engram parameters. During inference, it activates 12 billion parameters. It supports multimodal processing and features a context window of up to 1 million tokens, with a reported generation speed of around 200 tokens per second. These specifications suggest a design focused on efficiency and broad application.

    A key aspect of Hy4 flash's design is its reported cost-efficiency. Tencent’s pricing for the model is 0.5 yuan ($0.07) per million input tokens and 2 yuan ($0.3) per million output tokens, with cached input tokens costing 0.02 yuan ($0.003) per million. This pricing structure aims to make AI services more accessible for frequent use.

    According to Tencent’s internal WorkIndex benchmark, which evaluates common office tasks, Hy4 flash costs one-thirteenth as much as Hy4 preview and one-twentieth as much as Hy4 to complete a single task. Its cost is also stated to be around one-eighth that of mainstream lightweight models in the same category. These cost advantages, combined with its performance, make it suitable for high-frequency office tasks such as creating and editing presentations, document processing, spreadsheet repair, audio and video processing, and visual understanding. This suggests a shift in focus for AI models, where balancing performance with operational cost becomes as crucial as raw capability.

    The launch of Hy4 flash reflects a broader competitive shift in the AI industry. As daily token usage in China grows, tokens are increasingly becoming a key metric for AI adoption. Local governments are also introducing measures, from computing power and model vouchers to token-based financing, to make AI services more affordable. Tencent’s 50-trillion-token debut suggests a market demand for AI models that are not only capable but also affordable enough for frequent, everyday use. However, Tencent also notes that there is still a gap between strong first-day usage and long-term commercial value, indicating that sustained user value remains a key test.

    Share this article

    Original reporting by Tech NodeWe don't republish, read the full story →

    Related reading

    6 stories