← Volver a la portada
Asia · Mercados regionales · Yicai Global

DeepSeek Sets New Limits for Free Users Due to Insufficient Computing Power

Publicado por Yicai Global · China hace 9 h
DeepSeek Sets New Limits for Free Users Due to Insufficient Computing Power
DeepSeek Sets New Limits for Free Users Due to Insufficient Computing Power · Yicai Global

Resumen detallado

Many DeepSeek users have complained that the Chinese artificial intelligence giant has adjusted free usage limits for its app and web version because of “insufficient computing…

(Yicai) Nov. 17 -- Many DeepSeek users have complained that the Chinese artificial intelligence giant has adjusted free usage limits for its app and web version because of “insufficient computing power.”

The adjustments include daily usage quotas, with some users reporting online that they received a notice that “due to busy servers, your daily usage limit has been reached” after using DeepSeek for less than two hours. In addition, the maximum length of a single text input has been reduced to about 40,000 characters, while another notice said long conversations inactive for more than seven days are no longer supported, but some users faced the block after just 72 hours.

DeepSeek has not issued a new unified quota policy or any related announcement, nor has it disclosed whether it will lift the restrictions. However, the issue has fueled speculation that the Hangzhou-based company may introduce paid consumer plans.

One affected user said he was willing to pay to continue using the DeepSeek app, but the company has not yet offered a paid option. Others have suggested launching monthly, quarterly, or annual subscription plans like other large language model developers so that heavy users can pay for more stable service and higher usage quotas.

As the user base grows and computing resources remain limited, DeepSeek may need to trade off model training, paid application programming interface services, and free chat services, an industry insider told Yicai. By limiting free users’ long conversations, input length, and usage quotas, it can control token consumption, the person pointed out.

DeepSeek’s V4.1-Flash-1017 model handled 59.2 trillion tokens on API aggregation platform OpenRouter in the week ended Nov. 15, ranking first worldwide.

The changes may have limited impact on light AI users, but can significantly affect those who use AI for long periods. For example, some users rely on DeepSeek to revise papers, organize lengthy documents, or carry out role-playing…

The market…

Inicio del artículo, extracto de Yicai Global. Para leer el resto, abra el artículo en el sitio del medio.

Leer el artículo completo en Yicai Global Inicie sesión para guardar

MarketNews muestra el titular y el resumen publicados por el medio. El artículo completo se lee en el sitio del editor. (www.yicaiglobal.com)

También te puede interesar