DeepSeek Processes 200M Tokens for Just 5 RMB

huangyun_122 · x · 2026-07-07

A user tested DeepSeek for translation and vector database ingestion tasks, processing roughly 200 million tokens for only 5 RMB. This highlights DeepSeek's extremely low inference costs, making it highly suitable for large-scale text processing scenarios.

Original post →

More from Models

Models channel →