Google's open-source terminal agent for Gemini — 1M context, browser tools, surprisingly good. And surprisingly expensive at scale.
Gemini CLI's 1M-context appetite is its superpower and its cost trap — load a whole codebase, every turn re-reads it. Gemini 3 Pro is $4.80/$28.80 per 1M tokens. Heavy use = $300-500/month on Google direct.
Sign up (free, no card), then generate an sk-cc-… key from your dashboard.
Configure Gemini CLI with custom OpenAI-compatible endpoint.
# Gemini CLI doesn't natively support OpenAI-compat endpoints yet,
# so use a generic OpenAI-shape client instead — for example with curl:
curl https://claudeapi.cheap/api/proxy/v1/chat/completions \
-H "Authorization: Bearer sk-cc-your-key-here" \
-H "Content-Type: application/json" \
-d '{"model":"gemini-3-pro-preview","messages":[{"role":"user","content":"Hello"}]}'Gemini CLI keeps every feature. Only the API endpoint changed — your prompts, tools, and behaviour are unchanged.
Newest, best reasoning + multimodal
2M context — long-doc work
High-volume cheap loops
Not natively yet — Gemini CLI assumes Google's auth flow. Today the cleanest path is to use Gemini through any OpenAI-compatible client (LangChain, LlamaIndex, plain OpenAI Python SDK) pointed at our proxy. We're tracking native Gemini CLI support.
Streaming, tool calls, and multimodal are still partial on our Gemini path — text-only non-streaming is fully supported today. Streaming + tools are on the roadmap.
Vertex requires a Google Cloud project, billing setup, IAM, and per-project rate limits. We're: one key, no GCP setup, 80% off list, crypto pay. Same Gemini models.
Free Basic plan, $19 lifetime Pro. Crypto only. No subscription.
Get an API key — free