DeepSeek cost anomaly: 50x concurrent pricing could break all curves

xeophon · x · 2026-07-31

Addressing concerns about abnormally high cache write costs for the new DeepSeek model, a developer speculated that the token usage matches the previous version, suggesting a billing error.

They further guessed that if v4 flash is actually priced at 50 requests for $0.02, this billing model would completely break existing cost-performance benchmark curves.

Related event: DeepSeek-V4-Flash Benchmarks Impress but Face Pricing Controversy(4 posts)→

Original post →

More from Models

Models channel →