Gemini Quick Query

Use Gemini CLI for tasks where it excels over Claude/Codex.

When to Use Gemini

Use Case	Why Gemini
Real-time info	Google Search grounding - knows today's news/docs
Web research	Native URL fetching, summarizes pages
Quick checks	Flash 2.5 is very fast, free tier
Second opinion	Different perspective on architecture/approach
Large context	1M tokens for massive files/logs

Command Format

# Quick query (non-interactive)
gemini "your question here"

# With specific model
gemini -m gemini-2.5-flash "quick question"
gemini -m gemini-2.5-pro "complex analysis"

# Fetch and analyze URL
gemini "summarize https://example.com/docs"

# Output as JSON (for parsing)
gemini "query" --output-format json

Models

• gemini-2.5-flash - Fast, free tier, good for quick checks
• gemini-2.5-pro - Deeper reasoning, larger context

Use Case Examples

Real-time Information

gemini "what's new in Next.js 15.2 released this week?"
gemini "current best practices for React Server Components 2025"

Web Research

gemini "fetch https://docs.example.com/api and explain the auth flow"
gemini "summarize the changes in this PR: https://github.com/org/repo/pull/123"

Quick Second Opinion

gemini "is using zustand or jotai better for this use case: [brief description]"
gemini "review this approach: [paste snippet] - any red flags?"

Large File Analysis

# Gemini has 1M token context - good for massive logs
cat huge-log.txt | gemini "find the root cause of the OOM errors"

Execution

1. Determine if query benefits from Gemini's strengths (real-time info, web, speed)
2. Choose model: gemini-2.5-flash for speed, gemini-2.5-pro for depth
3. Run query with timeout: 60000 (1 min for flash, increase for pro)
4. Present response with model used

Output Format

**Model:** gemini-2.5-flash

[Gemini's response]

---
**Note:** [Any caveats about real-time data freshness if relevant]

Comparison

Feature	Gemini	Claude	Codex
Google Search grounding	Yes	No	No
Real-time web info	Yes	Limited	No
Speed (Flash)	Very fast	-	-
Free tier	Yes	No	No
Context window	1M	200k	200k

For Multimodal (Audio/Video/Images)

Use the ai-multimodal skill instead - it uses Gemini API directly with full multimodal support:

• Audio transcription (up to 9.5 hours)
• Video analysis (up to 6 hours)
• Image understanding, OCR
• PDF processing

---

Remember: Gemini for real-time info and speed, Claude for deep coding, Codex for complex debugging.