Closed Alpha Testing Phase #5

Phase #5 builds on the learnings from Phase #4 and extends the system into its most critical testing layers yet - covering node classification, light payment flows, and advanced L3 abstractions. With DevNet6 live, this phase brings deeper coordination across session flows, model diversity, and early metadata rollup logic.

This phase also marks the beginning of L3/Appchain/Rollup validation efforts, pushing toward the foundation of Cortensor’s future testnet stack. Participants play a critical role in pressure-testing this architecture while earning $COR based on active, reliable node operations.

Contest Rules

Stake Requirement

  • New participants must stake 5,000 $COR per node (Pool 3) to qualify.

Node Uptime & Participation

  • Maintain active, responsive nodes throughout the 30–40 day testing window.

  • Operate nodes capable of supporting expanded model types and task flow edge cases.

Bug Reporting Participants are expected to:

  • Report bugs, crashes, or inconsistencies to the team

  • Assist in testing and validating experimental logic where applicable

  • Provide feedback via Discord or direct discussion channels

Disqualification Rights Cortensor reserves the right to disqualify participants for:

  • Misreporting node status or uptime

  • Gaming staking or reward mechanics

  • Any behavior damaging to network health or testing integrity

Rewards

Base Reward

  • $20/month per node with qualifying uptime

Multi-Node Bonus

  • 3+ nodes → +10% bonus

  • 5+ nodes → +20% bonus

Role-Based Bonus

  • Earn +5% per Discord role (capped at 30%)

Node Level Bonus

  • We will announce the node level bonus soon. We'll have a clearer understanding as we continue testing.

Performance-Based Prizes

  • Nodes that meet both thresholds receive an additional $5:

Testing Focus

Here’s the updated version of your Phase #5 component list, adjusted to reflect current progress, future iteration focus, and recent changes:

CognitiveLevel (Node Categorization)

  • Miner-side logic live for node classification based on performance

  • Enables adaptive routing and task-type optimization

  • Validator-side scoring active for SLA enforcement and history tracking

  • Upgrade/downgrade logic kicks in after 10+ verified tasks

  • More tuning expected on tier thresholds and fallback behavior

Validator v1 – Foundation for Session Verification

  • Tracks per-task result across sessions and miners

  • Early PoI and PoUW enforcement via task checkpoints

  • Light logic in place for runtime validation (not final scoring)

  • Sets the groundwork for decentralized validator rotation and rewards

SessionPayment – Light Payment Flow Testing

  • Event-based payout system now functional on Sepolia

  • Supports real-time credit allocation to miners upon task completion

  • Loosely coupled architecture tested, with contract-to-contract flow planned

  • COR token used for all session billing & miner compensation

User Task Mode – Model Diversity Expansion

  • Dedicated Node: Supports multiple models via config (e.g., DeepSeek-R1, LLaMA 3.1)

  • Ephemeral Node: Starting work on longer reservation for model-specific sessions

  • More iteration to ensure model switching, routing, and queue stability

L3 Layer – Appchain & Rollup Simulation

  • COR-as-gas active on Testnet-1 with oracles + miners

  • L3 explorer and bridge deployed for testing full stack

  • Testing recursive inference flow, metadata rollup, and multi-session stack orchestration

  • Dashboard + payment UX refactor in progress for full L3 compatibility

(Removed for Phase #5) App Registry (WIP)

  • Paused for now to focus on core task and SLA infrastructure

  • Will resume later to support agent-level accounting, traffic control, and reward tracking

Upcoming Public Hackathon Preview

  • Duration: 4 weeks

  • Incentive: Double rewards

  • Theme: Real-world app integrations and AI agent tools

  • Audience: Open to devs, builders, and contributors

  • Launch: Immediately after Phase #5 concludes

Key Outcomes

  • Confirm miner-level categorization and routing logic across sessions

  • Validate event-driven payment logic with testnet COR

  • Support dynamic model handling in open User Task Mode

  • Simulate metadata rollup and composability under L3 architecture

  • Prepare infrastructure for Appchain/Agent Registry design

  • Lay foundational testing for Cortensor's eventual testnet

Join the Contest

Push the boundaries of decentralized AI. Test advanced logic, validate core systems, and earn $COR while shaping the infrastructure behind Cortensor’s future stack.

Join Cortensor Discord: discord.gg/cortensor

PreviousClosed Alpha Testing Phase #4NextDiscord Roles & Mainnet Privileges

Last updated