Using Amazon Bedrock with AWS Free Tier for the 10,000 AIdeas Competition 🛠️

Published: (January 6, 2026 at 11:00 AM EST)
5 min read
Source: Dev.to

Source: Dev.to

What’s This Hackathon All About?

The Global 10,000 AIdeas Competition is one of the largest developer contests in AWS history. It invites developers worldwide to build innovative AI applications using AWS tools.

  • Prizes: $250 k in cash, plus recognition at AWS re:Invent 2026 and across AWS channels.
  • Tracks:
    1. Workplace efficiency
    2. Daily‑life enhancement
    3. Commercial solutions
    4. Social impact
    5. Creative expression

Your first submission is a written pitch (no code required) that explains what you want to build and why it matters.

If you become a semi‑finalist, you’ll develop the app. The finished solution must:

  • Use the agentic IDE Kiro for at least part of the development.
  • Stay within AWS Free Tier limits.
  • Be completely original and not yet published.

For more details and FAQs, see the official announcement post.


The New Free Tier: Everything You Need to Know

AWS recently revamped the Free Tier for new accounts created after July 15 2025. Here’s the breakdown:

CreditHow to EarnTotal
$100Sign‑up bonus
Up to $100Complete five onboarding activities (one is a Bedrock playground task)Up to $200
  • Duration: 6 months or until you exhaust your credits.
  • Bedrock activity: Using Amazon Bedrock in the console (Bedrock Playground) earns you part of the extra credit.

Note: Amazon Bedrock is not part of the “Always Free” tier. You’ll be charged against your free credits for any Bedrock usage.


How Amazon Bedrock Pricing Works

Bedrock pricing is token‑based (or per‑image for vision models). Understanding the model helps you predict costs accurately.

CategoryPricing Model
Text modelsPay per input + output token
Image modelsPay per image generated
EmbeddingsPay per input token

Bedrock offers three pricing tiers:

  1. On‑Demand – Pay‑as‑you‑go (recommended for the competition).
  2. Batch Mode – 50 % discount for non‑urgent workloads.
  3. Provisioned Throughput – Reserved capacity (overkill for this hackathon).

For full details, visit the official Amazon Bedrock pricing page.


Which Models Should You Use?

Not all models cost the same. Choosing the right one can stretch your $200 credits dramatically.

Budget‑Friendly Champions 🏆

All prices are for the us‑east‑1 (N. Virginia) region.

ModelInput (per 1 K tokens)Output (per 1 K tokens)Why Choose It
Amazon Nova Micro$0.000035$0.00014Cheapest Amazon model; great for simple routing or classification.
Amazon Nova Lite$0.00006$0.00024Still very affordable; solid reasoning capabilities. My pick for most hackathon use cases.
Meta Llama 3.3 70B$0.00072$0.00072Good price‑to‑performance; strong reasoning and tool use; flat pricing simplifies cost prediction.

Where to Splurge (Strategically) 🤑

ModelInput (per 1 K tokens)Output (per 1 K tokens)When to Use It
Claude Sonnet 4.5$0.0033$0.0165High‑quality, complex reasoning; ideal for demo/presentation where you want to wow the judges.
Claude Haiku 4.5$0.0011$0.0055Faster and cheaper than Sonnet; good middle ground for production‑like scenarios. My pick if you want to splurge modestly.

Further reading: The deep‑dive article “Claude Haiku 4.5” by Anthropic explains why this model is both performant and affordable.


The Power Move: Intelligent Prompt Routing

Pro tip: Use Bedrock’s Intelligent Prompt Routing to automatically send simple queries to cheap models (e.g., Nova Lite or Haiku) and route complex queries to more capable models (e.g., Sonnet or Nova Pro).

  • Result: Up to 30 % cost reduction.
  • Example: In a customer‑service chatbot, a question like “What are your hours?” is handled by Nova Lite, while a multi‑step troubleshooting request is escalated to Claude Sonnet.

Implement routing logic in your application code (or via Bedrock’s built‑in feature) to maximize credit efficiency.


Quick Checklist for the Competition

  • Create a new AWS account (post‑July 15 2025) to qualify for the revamped Free Tier.
  • Earn the full $200 credit by completing the onboarding activities, including the Bedrock playground task.
  • Select cost‑effective models (Nova Lite, Llama 3.3) for the bulk of your workload.
  • Reserve higher‑tier models (Claude Sonnet/Haiku) for demo‑critical paths.
  • Enable Intelligent Prompt Routing to automatically optimize per‑request costs.
  • Monitor usage in the AWS Billing console to stay within your credit limit.

With this strategy, you can build a competitive AI application, showcase impressive capabilities, and keep the AWS bill at $0. Happy hacking!


Prompt Routing & Cost‑Effective Tips for AWS Bedrock

  • “I need a quick answer about pricing” → routed to a cheap model.
  • “I need help troubleshooting this complex integration issue” → routed to the smart (expensive) model.

There’s a cost associated with this feature—$1 per 1,000 requests—but it can be worth experimenting with! For more details, see the Prompt Routing documentation.


Maximizing Your Free Credits: Battle‑Tested Tips

1. Start with Nova Lite or Nova Micro

Build your application with these models first. Get the logic working, nail your tool integrations, and perfect your prompts. Upgrade to more expensive models only when you’re confident the app works well.

2. Use Batch Mode for Testing

If you have large test suites or want to evaluate multiple prompt variations, try Batch mode for a 50 % discount compared to on‑demand. It isn’t real‑time, but it’s great for development and testing.

3. Optimize Your Prompts

Shorter prompts = fewer tokens = lower costs. Be concise in system prompts and examples—every word counts.

4. Implement Prompt Caching

Avoid redundant API calls. If your agent asks the same question multiple times, cache the response. This feature isn’t available for all models, so double‑check the Prompt Caching documentation.

5. Monitor Costs in Real‑Time

  1. Open AWS Cost Explorer → filter by:
    • Granularity: Daily
    • Dimension: Usage Type (shows which models are costing you)
    • Service: Bedrock
  2. Check this daily during development. You don’t want to burn through $150 of credits on day 3 and only notice it on day 10.

6. Set Up Budget Alerts

  1. Go to AWS Budgets → create a budget.
  2. For the competition, simulate the burn rate of your free‑tier credits:
    • Budget: $200
    • Alerts:
      • 50 % ($100 spent)
      • 75 % ($150 spent)
      • 90 % ($180 spent)

⚡️ Bonus: Setting up a cost budget is one of the five onboarding tasks for the new Free Tier, unlocking an extra $20 / $100 in AWS credits.

7. Use the Playground First

The Bedrock Playground lets you test prompts without writing code. It costs the same but is much faster for iteration. Plus, using it earns you $20 in credits!


Regional Considerations

Not all models are available in every region. The most model variety is in:

RegionDescription
us-east-1 (N. Virginia)Best option for maximum model access
us-west-2 (Oregon)Good alternative with most models

Stick to these regions for the competition to avoid “model not available” headaches.


Final Thoughts

If you’re a visual/video learner, we have FREE live learning sessions this week from AWS experts covering how to use the AWS Free Tier.

AudienceDate & TimeRegister
AmericasJan 6, 12 pm PT (GMT‑8)Register
EMEAJan 7, 12 pm London Time (GMT+0)Register
APJ (Japan, China)Jan 7, 12 pm AEDT (GMT+11)Register

Now go build something cool—and set up those budget alerts first. Your wallet will thank you! 😊

Additional Resources 📚

Back to Blog

Related posts

Read more »

Rapg: TUI-based Secret Manager

We've all been there. You join a new project, and the first thing you hear is: > 'Check the pinned message in Slack for the .env file.' Or you have several .env...

Technology is an Enabler, not a Saviour

Why clarity of thinking matters more than the tools you use Technology is often treated as a magic switch—flip it on, and everything improves. New software, pl...