Mistral for fast production work.
Add Mistral to your model mix without another app-level client. Your key passes through one endpoint, and budget caps plus usage analytics wrap every request.
$129/month SaaS. Bring your own model keys. No inference markup.
Three steps to connect.
Pass Mistral through today
Use Mistral models through OpenRouter-backed access on your own key. Native Mistral key storage is a later direct-provider step, not a requirement for this setup.
Point your SDK
Set the OpenAI-compatible base URL to https://api.proxyllm.ai/v1 and authenticate with your ProxyLLM key.
Put Mistral on high-volume work
Run mistral-large and Mixtral on chat, extraction, and multilingual tasks, then compare latency and cost in the request logs.
Mistral from one endpoint.
Use mistralai/ model names through the OpenAI-compatible gateway on your own key.
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.proxyllm.ai/v1",
apiKey: "pk_live_...",
});
const r = await client.chat.completions.create({
model: "mistralai/mistral-large",
messages: [{ role: "user", content: "Extract account fields as JSON." }],
});Run your AI workloads on your ChatGPT subscription.
ProxyLLM runs OpenAI's Codex for you, signed in with your own ChatGPT account. Your apps call one OpenAI-compatible endpoint and the work bills to your flat plan instead of per-token API pricing.
Benchmark Mistral with real traffic.
Compare quality, latency, and cost beside OpenAI, Claude, Gemini, and Llama in one request log. OpenAI-bound work can run on Codex Hosted with a flat ChatGPT subscription; Mistral passes through on your key.