Running Mirofish on Infernex GEO - 95% cheaper than Open AI and fully compatible with Open AI just 2 like of code to change et voilà

[mercuriumbsc]

Hey MiroFish community 👋

I've been experimenting with running MiroFish simulations using Infernex Geo
as the LLM backend instead of OpenAI, and the cost savings are significant enough
that I wanted to share the setup here.

Why it matters for MiroFish users

MiroFish can spawn thousands to millions of agents, each making LLM calls.
At OpenAI pricing, a large simulation can get expensive fast.
Infernex offers tokens from $0.05/M — up to 95% cheaper — with an
OpenAI-compatible API, so the swap is literally a one-line config change.

Config change (.env)

# Before (OpenAI)
OPENAI_API_KEY=sk-...
OPENAI_BASE_URL=https://api.openai.com/v1

# After (Infernex Geo)
OPENAI_API_KEY=your-infernex-key
OPENAI_BASE_URL=https://geo.infernex.ai/v1

That's it. No code changes needed.

What I tested

• ✅ 1,000 agent simulation — worked perfectly
• ✅ GraphRAG construction — no issues
• ✅ ReportAgent output — same quality
• ⚡ Latency: <5ms, multi-region routing via Infernex Geo

Free tier available

Infernex has a free tier if you want to try before committing:
👉 https://infernex.ai/signup

Has anyone else experimented with alternative LLM backends for MiroFish?
Would love to hear what setups the community is using.