FRAUD!!! Fake caching – you pay ~10x more for tokens than necessary
Ollama Cloud pretends to cache but doesn't. That's fraud.
Ollama Cloud claims to cache prompts and pass the savings on to subscribers. It doesn't. Independent tests show clearly and repeatedly that no caching is happening – despite the fact that they market it. The result: every single one of us is paying roughly 10x more for our LLM tokens than we actually have to.
This is not a small discrepancy or a "coming soon" feature. It's a core part of their monthly subscription pitch, and it simply doesn't work. They keep your money and keep the token savings for themselves.
This is an absolute disgrace – and frankly, it's fraud. Paying a monthly subscription and then being quietly overcharged on every token because a promised feature silently doesn't do anything is exactly the kind of behavior that should be called out publicly.
Check it yourself before you spend another cent. Run a simple multi-turn request and watch the token billing. You'll see it immediately. Then decide whether you want to continue supporting a company that deceitfully rips you off and cheats you.






