How do I use Gemini in Hermes Agent?
Hermes supports Gemini with a Google AI Studio API key (its gemini provider), or with a standardcompute entry under providers: in ~/.hermes/config.yaml (api https://api.stdcmpt.com/v1) and google/gemini-3.8-flash as the default model. The second runs on a plan from $19 a month that returns HTTP 402 to new requests once the compute budget is used.
# Gemini 3.8 Flash by default, 3.1 Pro for hard steps
# Put STANDARDCOMPUTE_API_KEY=YOUR_API_KEY in ~/.hermes/.env
providers:
standardcompute:
api: https://api.stdcmpt.com/v1
key_env: STANDARDCOMPUTE_API_KEY
transport: chat_completions
default_model: google/gemini-3.8-flash
models:
google/gemini-3.8-flash: { context_length: 400000, supports_vision: true }
google/gemini-3.1-pro-preview: { context_length: 400000, supports_vision: true }
model:
provider: custom:standardcompute
default: google/gemini-3.8-flash
This is Hermes' named-provider format over Chat Completions; Hermes' own Gemini provider talks to Google's native API instead.
What happened to Gemini sign-in in Hermes
Hermes removed its google-gemini-cli and google-antigravity sign-in providers in v2026.7.1, after maintainers identified account-ban risk for third-party use of Google's Code Assist sign-in. The supported Google route is now the gemini provider with a GEMINI_API_KEY or GOOGLE_API_KEY. Hermes' Gemini guide recommends a billing-enabled project, since the free tier is too small for long agent sessions.
What a Hermes day on Gemini costs
| Per day | 3.8 Flash, 2026 | 3.8 Flash, 2027 | 3.1 Pro |
|---|---|---|---|
| Typical day | $1.54 | $3.08 | $4.19 |
| Busy day | $16 | $31 | $43 |
Gemini API paid tier per million tokens: 3.8 Flash $0.75 input, $0.075 cached, $3.75 output through December 31, 2026, then $1.50, $0.15, $7.50; 3.1 Pro $2, $0.20, $12 up to 200K. Source. The Agent Burn Index measures coding agents such as OpenCode, Cline and Kilo; Hermes Agent is not broken out, so these figures use its median and 90th-percentile coding-agent day (September 13 to October 12, 2026) as a stand-in. Priced at the list rates above; your own days will differ.
What works with Gemini in Hermes
- Tools, skills and MCP work over Chat Completions; both models read images, so Hermes sends screenshots directly.
- Hermes sends reasoning_effort medium unless you change it; Gemini cannot switch reasoning off, so /reasoning none runs at low.
/model custom:standardcompute:google/gemini-3.1-pro-previewmoves a hard job to 3.1 Pro.
The catch
With your own AI Studio key, Hermes' gemini provider pays Google directly over Google's native API, with no gateway in between. For a Hermes that only ever uses Gemini, that is the shorter path.
Where Standard Compute fits
A monthly cap for an agent that keeps working overnight: at 100% of the budget, requests stop with HTTP 402 instead of running up a Google bill.
FAQ
Does Hermes' gemini provider use the same endpoint as this config?
No. The built-in provider calls Google's native generateContent API; this config uses Chat Completions at api.stdcmpt.com. Both reach the same Gemini models.