Model Selection

Choose the right Gemini model for your task.

Heads up — backend matters. Model selection below applies to the default Gemini CLI backend. Google retires the Gemini CLI on 2026-06-18 for free/Pro/Ultra tiers; the experimental Antigravity CLI (agy) backend runs print-mode as Gemini 3.5 Flash only and ignores model selection. See the Antigravity migration guide for details.

Available Models

Gemini-2.5-pro

• Best for: Complex analysis, large codebases
• Context: 2M tokens
• Use when: Analyzing entire projects, architectural reviews, stronger reasoning

Gemini-2.5-flash

• Best for: Quick responses, routine tasks
• Context: 1M tokens
• Use when: Fast code reviews, Analyzing entire projects, simple explanations

Setting Models

You need use natural language: "...using gemini flash"

You can also append with '-m' or ask specifically with

In Configuration

{
  "mcpServers": {
    "gemini-cli": {
      "command": "gemini-mcp",
      "env": {
        "GEMINI_MODEL": "gemini-1.5-flash"
      }
    }
  }
}

Per Request (Coming Soon)

/gemini-cli:analyze --model=flash @file.js quick review

Model Comparison

Model	Speed	Context	Best Use Case
Pro	Slower	2M tokens	big ideas
Flash	Fast	1M tokens	quick, specific changes

Cost Optimization

1. Start with Flash for most tasks
2. Use Pro only when you need the full context
3. Flash-8B for simple, repetitive tasks

Token Limits

• Pro: ~2 million tokens (~500k lines of code)
• Flash: ~1 million tokens (~250k lines of code)
• Flash-8B: ~1 million tokens (~250k lines of code)

Recommendations

• Code Review: Flash
• Architecture Analysis: Pro
• Quick Fixes: Flash-8B
• Documentation: Flash
• Security Audit: Pro