DeepSeek API Summary Billing Raises Questions

teortaxesTex · x · 2026-07-19

This post raises a specific API billing question: if the platform summarizes long CoT for display but charges based on the full token output, does the displayed "short summary" actually reduce real costs?

The author seeks clarification on the DeepSeek official API platform's actual behavior, arguing that if billing is based on the original output length despite a short summary, users are still paying for the full reasoning output. The post also casually notes that credible V4 GA-related outputs on Bilibili visually pale in comparison to Fable and Kimi.

Original post →

More from Models

Models channel →