Developers
Get Started
Model Forge Competition

Model Forge Competition

Build and deploy AI models to compete in Allora's hackathon-style prediction competition. Submit inferences, earn rewards based on accuracy.

Competition Overview

  • Open-source hackathon for AI model development
  • Accuracy-based scoring against ground truth data
  • Network rewards for top-performing models
  • Real-time leaderboards and performance metrics

Available Topics

You can develop and deploy AI models using any of these active prediction topics on Allora Testnet:

Topic IDMetadataDefault Arg
1ETH 10min PredictionETH
2ETH 24h PredictionETH
3BTC 10min PredictionBTC
4BTC 24h PredictionBTC
5SOL 10min PredictionSOL
6SOL 24h PredictionSOL
7ETH 20min PredictionETH
8BNB 20min PredictionBNB
9ARB 20min PredictionARB
10Memecoin 1h PredictionTOKEN_FROM_API
11US Presidential Election 2024 - Winning PartyR
13ETH 5min PredictionETH
14BTC 5min PredictionBTC
15ETH 5min Volatility PredictionETH
17ETH 8h PredictionETH
18BTC 8h PredictionBTC
37SOL/USD - 5min Price PredictionSOL
38SOL/USD - 8h Price PredictionSOL
41ETH/USD - 8h Price PredictionETH
42BTC/USD - 8h Price PredictionBTC
475 min BTC Price PredictionBTC
506h BTC Volatility PredictionBTC
561 hour BERA/USD Log-Return PredictionBERA
588 hour SOL/USD Log-Return PredictionSOL
6024 hour XAU/USD Log-Return PredictionXAU
611 day BTC/USD Log-Return PredictionBTC
621 day SOL/USD Log-Return PredictionSOL
631 day ETH/USD Log-Return PredictionETH

Topic Selection: Choose topics that align with your model's strengths and interests. Each topic has different prediction challenges and reward structures.

Quick Registration

Create Wallet & Get Funds

# Install CLI (if not done)
curl -sSL https://raw.githubusercontent.com/allora-network/allora-chain/main/install.sh | bash -s -- v0.8.0
 
# Create wallet
allorad keys add <key-name>
 
# Get testnet funds
# Visit: https://faucet.testnet.allora.network/

Register for Competition

  1. Sign up: Visit the Model Forge Registration (opens in a new tab) page
  2. Submit wallet address from step 1
  3. Join Discord (optional): Allora Discord (opens in a new tab) for exclusive support channels

Connect & Verify

  1. Install Keplr: Chrome Extension (opens in a new tab)
  2. Connect wallet to Allora Forge site

connect-wallet

  1. Verify eligibility - You should see "eligible" status in the top left corner

eligible

Congratulations! You're now registered for the Allora Model Forge Competition.

Model Development

Allora MDK Framework

Use the Allora Model Development Kit (opens in a new tab) for rapid model development:

Features:

  • Multiple regression strategies (Linear, SVR, Ridge, etc.)
  • Optimized for price/volatility/volume predictions
  • Easy customization and deployment
  • Comprehensive documentation

Quick Start:

git clone https://github.com/allora-network/allora-mdk
cd allora-mdk
# Follow setup instructions in repository

Topic Structure

Understanding topic parameters for model development:

# Get topic details
allorad query emissions topic <topic-id> --node https://allora-rpc.testnet.allora.network/
 
# Example for ETH 10min prediction (Topic 1)
allorad query emissions topic 1 --node https://allora-rpc.testnet.allora.network/

Key Fields:

  • Target Variable: Asset to predict (ETH, BTC, SOL, etc.)
  • Epoch Length: Submission frequency (varies by topic)
  • Loss Function: Depends on topic type
  • Metadata: Prediction timeframe and details

Performance Monitoring

Live Leaderboards

Track your model's performance on the Allora Forge Platform (opens in a new tab):

Connect your wallet to the Allora Forge site to view active competitions and see your ranking across different prediction topics.

Allora Explorer Metrics

View detailed performance:

  1. Connect wallet to Allora Explorer (opens in a new tab)

connect-wallet

  1. Click "Your Topics" to see your submissions

your-topics

  1. Select topic for detailed metrics

topic-metrics

Key Metrics:

  • Your Score: Overall performance ranking
  • ALLO Earned: Rewards based on accuracy
  • Inference History: Track prediction accuracy over time

Development Tips

Model Optimization:

  • Understand the loss function for your chosen topic
  • Test with historical data before deployment
  • Monitor epoch timing for submission windows
  • Use MDK framework for faster iteration

Best Practices:

  • Data quality: Ensure clean, relevant training data
  • Model validation: Cross-validate before topic submission
  • Resource management: Optimize for topic-specific epoch intervals
  • Community engagement: Join Discord for tips and support

Getting Help

Next Steps

  • Start Building: Clone the MDK repository and begin model development
  • Join Community: Connect with other participants on Discord
  • Monitor Performance: Use Explorer to track your model's accuracy
  • Scale Up: Deploy multiple models or optimize existing ones