Supported Models & Reasoning Controls

gemini-relay connects your AI agents directly to Google's latest Gemini 3.8 and Gemini 3.1 model family via the Antigravity CLI (agy) backend.

---

🌟 Available Model Family

Model Identifier	Alias	Best For	Reasoning Depth	Speed
gemini-3.8-flash-high	flash	Default choice. Code review, large refactoring, Q&A, and fast planning.	Deep	Ultra-Fast (~2-4s)
gemini-3.8-flash-medium	-	Everyday tasks with balanced reasoning depth.	Medium	Ultra-Fast (~1-3s)
gemini-3.8-flash-low	-	High-throughput summaries, syntax checks, formatting.	Low	Instant (<1s)
gemini-3.1-pro-high	pro	Deep architectural reasoning. Algorithm design, math proofs, security boundary audits.	Maximum Flagship	Moderate (~5-10s)
gemini-3.1-pro-low	-	Pro-grade knowledge retrieval with minimal thinking delay.	Low	Fast (~2-4s)
gemini-3.7-flash-high	-	Previous generation high-reasoning flash model.	High	Fast

---

🧠 Reasoning Effort Depth

You can control thinking token allocation dynamically on any query by passing the effort parameter:

{
  "name": "ask-gemini",
  "arguments": {
    "prompt": "Audit @src/security.ts for subtle timing attacks and side-channel leakage.",
    "model": "gemini-3.1-pro-high",
    "effort": "high",
    "includeUsage": true
  }
}

• effort: "high": Allocates maximum reasoning tokens for rigorous edge-case analysis, distributed systems design, or security audits.
• effort: "medium": Standard balanced reasoning for refactoring and feature implementation.
• effort: "low": Minimal thinking overhead for instant summaries, code classification, and translation.

---

📋 Discovering Available Models

You can ask your agent to inspect available models at any time using the gemini-models tool:

"Run gemini-models to see what Gemini models and reasoning options are available."