agentos.guide › blog
Hermes Agent Gemini 4 is a quick upgrade for your Agent OS: add a Google API key, run hermes model, choose Google AI Studio and pick Gemini 4 Hydrogen, Argon or Carbon.
Book a free SEO strategy session →
This guide gives you the setup, a sensible model split and prompts to put Gemini 4 to work in your Agent OS.
If you would like agents working on your website's growth, book a free SEO strategy session.
Facts are from the Hermes Gemini guide and Google's Gemini 4 pages.
# 1. add your key (GEMINI_API_KEY also works)
echo "GOOGLE_API_KEY=your-key-here" >> ~/.hermes/.env
# 2. pick the provider and model
hermes model
# More providers... -> Google AI Studio -> gemini-4-hydrogen
# 3. check it, then chat
hermes doctor
hermes chat
You need a key from Google AI Studio and billing switched on; the free allowance runs out fast for an always-on agent.
| Job | Model | Effort |
|---|---|---|
| Chat replies and quick tasks | gemini-4-hydrogen | medium |
| Long reports and big documents | gemini-4-hydrogen | xhigh |
| Plans and careful analysis | gemini-4-argon | high |
| The hardest problems | gemini-4-carbon | high |
For how Gemini compares with other models in Hermes, see our guide to the best LLMs for Hermes Agent.
🔥 Want Gemini 4 wired into your Agent OS today?
Inside the AI Profit Boardroom you get step-by-step video trainings, the installable Agent OS, four coaching calls a week and 3,400+ members testing the newest models on real business work.
For quick replies use gemini-4-hydrogen. If I say "think hard",
switch with /model gemini-4-carbon for that task, then switch back
to gemini-4-hydrogen when you are done. Tell me which model you used.
Set reasoning to xhigh. Using everything in [folder], write a full
report with an executive summary, findings, evidence and next steps.
Do not stop early; finish the whole report in one reply.
At xhigh, Hermes turns on long decode, which lets Gemini 4 keep writing past the normal output limit.
Avoid this setting for private customer data, because Google stores long-decode requests for a period.
📞 Want agents working on your search traffic?
Book a free SEO strategy session and we will show you how Hermes and Gemini 4 fit into your SEO plan.
Book a free SEO strategy session →
Add claude-opus-5-5 as a fallback provider after gemini-4-hydrogen
using hermes fallback, then run hermes fallback check --live and show
me the result.
Hermes switches to the backup automatically if Gemini is busy or blocks a request, per its fallback docs.
Every morning at 8am, check my inbox and calendar, then post a
briefing in Telegram: today's meetings, the three most urgent
emails and one suggested priority. Use gemini-4-hydrogen at medium.
Scheduled briefings are a good first job for Gemini 4 in an Agent OS: short, useful every day and cheap on Hydrogen.
hermes gateway setup
hermes gateway start
Gemini works on every Hermes chat platform, including Telegram, Discord, Slack and WhatsApp, per the Gemini guide.
Newer Google keys start with AQ. for both AI Studio and Vertex AI express, so Hermes uses the base URL to decide where to send requests.
For an express key, add GEMINI_BASE_URL=https://aiplatform.googleapis.com to ~/.hermes/.env and restart.
Gemini 4 Argon only runs in Google's global location, so it needs the longer project-scoped URL ending in /locations/global.
hermes model.hermes model.Run hermes doctor after any change; it checks your key and provider settings.
Once a week, tell me which Gemini 4 model handled each type of
task, roughly how much thinking each used, and which tasks could
move down to gemini-4-hydrogen without losing quality.
Hydrogen costs less per word than Argon and Carbon, per Google's Hydrogen notes, so moving routine jobs down is the easiest saving.
Book a free SEO strategy session →
All three can read up to about a million tokens at once, enough for a large folder of documents.
Use gemini-4-carbon to write a plan for [project] with numbered
steps. Then switch to gemini-4-hydrogen with /model and carry out
each step, checking it against the plan. Report progress after
every three steps.
Carbon does the careful thinking once, and Hydrogen does the volume work at a lower price.
Hermes ships updates often, listed on its release page, and recent ones improved Gemini handling, including moving Gemini safety blocks to your backup model.
Update regularly and run hermes doctor after each update.
Set Hydrogen as your default, keep Carbon for hard jobs and always add a backup.
Members of the AI Profit Boardroom get the Agent OS on Hermes with Gemini and Claude routing built in.
Book a free SEO strategy session →
Add GOOGLE_API_KEY to ~/.hermes/.env, run hermes model, choose Google AI Studio and pick a Gemini 4 model.
Hydrogen for most jobs, Argon for plans and analysis, Carbon for the hardest problems.
Set reasoning effort to xhigh or max, which turns on long decode in Hermes.
Set a fallback with hermes fallback and test it with hermes fallback check --live.
Default to Hydrogen, add a backup and Hermes Agent Gemini 4 will run your Agent OS well; for search growth, book a free SEO strategy session.
About Julian
I'm Julian Goldie, an AI entrepreneur, SEO expert and the founder of the AI Profit Boardroom, which has 3,400+ members.
I help business owners scale with AI agents, automation and SEO.
My YouTube channel has 400,000+ subscribers, and I run Goldie Agency, a seven-figure SEO agency.
→ The Best Hermes Agent Model, And The Five Model Lanes I Route Inside My Agent OS
→ Hermes Subagent Model: The Quick Start, And How To Set Worker Models In Your Agent OS
→ Hermes Agent Backend: The Quick-Start, Every Option, And Where Each One Fits In An Agent OS
→ Gemini 4 Hydrogen vs Argon: 3 Prompts to Test Both
The installable system behind every guide on this site — dashboard, agents, memory and pipelines — updated as new tools land, with 3,400+ members and four coaching calls a week.
Join the AI Profit Boardroom →📺 Video notes + links to the tools 👉
🎥 Learn how I make these videos 👉