Stop Using OpenAI’s API, Use This One Instead

The recent drama at OpenAI, following the abrupt dismissal and reinstatement of Sam Altman, has shown to us how imprudent it is to rely on only one LLM provider. Everything goes fast in the AI space. What if your LLM provider implodes in the next few weeks ? It almost happened for OpenAI right ?

In this article, I will explain why you should drop OpenAI’s API, and use Mistral’s API, or at least use both in tandem.

Mistral AI, the parisian startup poised to dethrone OpenAI, has been releasing banger LLM models recently. And they do it without fuss, without faking demo videos like Google, just dropping a torrent link on X (formerly Twitter).

Mistral now has three models available on their API: mistral-tiny, mistral-small, and mistral-medium.

Mistral small is a high performing LLM on par with GPT-3.5. If you use it, you get GPT-3.5’s capabilities at a lower cost. Mistral small costs 0.6€/1M input tokens, and 1.8€/1M output tokens. GPT-3.5-turbo costs 0.9€/1M input tokens and 1.8€/1M output tokens. Mistral small is thus cheaper than…