How to Use Gemini 3.1 Flash for Fast, Cheap AI Tasks

Practical Implementation

Example 1: Email Classification Pipeline

import google.generativeai as genai

genai.configure(api_key="your-key")
model = genai.GenerativeModel('gemini-3.1-flash')

def classify_email(email_text: str) -> dict:
    prompt = f"""Classify this email into exactly one category:
    - support_request
    - billing_inquiry
    - feature_request
    - bug_report
    - spam
    - other

    Also extract: urgency (low/medium/high), sentiment (positive/neutral/negative)

    Email: {email_text}

    Respond in JSON only."""

    response = model.generate_content(prompt)
    return json.loads(response.text)

# Process 10,000 emails for ~$0.75

Example 2: Bulk Content Summarization

async def summarize_articles(articles: list[str]) -> list[str]:
    """Summarize 1000 articles in parallel using Flash"""
    tasks = []
    for article in articles:
        prompt = f"Summarize in 3 bullet points:\n{article}"
        tasks.append(model.generate_content_async(prompt))

    responses = await asyncio.gather(*tasks)
    return [r.text for r in responses]

Example 3: Smart Routing

def smart_route(query: str) -> str:
    """Use Flash to classify, then route to the right model"""
    # Step 1: Flash classifies the task (cheap)
    complexity = classify_complexity(query)  # Uses Flash

    # Step 2: Route to appropriate model
    if complexity == "simple":
        return call_flash(query)      # $0.075/M tokens
    elif complexity == "moderate":
        return call_sonnet(query)     # $3/M tokens
    else:
        return call_opus(query)       # $15/M tokens

Flash-Specific Optimization Tips

1. Keep prompts short — Flash works best with concise instructions
2. Use structured output — JSON mode reduces parsing errors
3. Batch requests — process multiple items per call when possible
4. Cache common prompts — use Google’s context caching feature
5. Set temperature low — 0.0-0.3 for classification, extraction, and formatting tasks

People Also Ask

Is Gemini Flash good enough for production?

For the right tasks, absolutely. Classification, extraction, summarization, and formatting tasks run well on Flash. Don’t use it for tasks where quality errors have high consequences.

How does Flash compare to Claude Haiku?

Flash is cheaper per token. Haiku has slightly better quality for instruction-following tasks. Both are excellent for high-volume, simple tasks. Test both on your specific use case.

Can Flash handle long documents?

Yes — Flash supports the same 1M token context window as Gemini Pro. It’s excellent for processing long documents when you need extraction or summarization rather than deep analysis.

Want to skip months of trial and error? We’ve distilled thousands of hours of prompt engineering into ready-to-use prompt packs that deliver results on day one. Our packs at wowhow.cloud include battle-tested prompts for marketing, coding, business, writing, and more — each one refined until it consistently produces professional-grade output.

Blog reader exclusive: Use code BLOGREADER20 for 20% off your entire cart. No minimum, no catch.

Browse Prompt Packs →