Ask HN: Why is cache for DeepSeek-v4 cheapest on Vercel AI Gateway?

Do they charge below their cost? Or do they run their own cache?

Vercel cache read: $0.01/M for flash, $0.14/M for pro.

Deepseek and OpenRouter cache read: $0.028/M for flash, $0.145 for pro.

That's a 64% discount on Vercel for flash, and a 3% discount for pro.

Links: Vercel - https://vercel.com/ai-gateway/models OpenRouter - https://openrouter.ai/deepseek/deepseek-v4-flash