VergeX is partnering with Z.ai to launch the GLM-5 Trading Odyssey, a live trading initiative designed to explore how AI agents perform in real markets.

Unlike simulations or paper trading environments, this initiative operates directly in live market conditions.

Participants will:

• Build trading agents using GLM-5

• Deploy and run strategies on Vergex

• Execute trades in real markets

• Compete based on actual performance

All participants operate on the same model.

Same model. Different outcomes.

Vergex and Z.ai will provide GLM-5 token credits to support participation.

Credits are structured to encourage real usage:

• All users (new and existing) receive 5M GLM-5 tokens at campaign start or upon registration

To retain credits, users must:

• Connect an exchange account

• Run at least one agent

Credits expire within 48 hours if these conditions are not met.

Participants can earn additional credits by:

• Sharing results (PnL / strategies)

• Contributing active strategies

Top-performing participants will receive:

• Additional credits

• Official exposure by Vergex × Z.ai

• A connected exchange account is required

• Minimum balance: ≥50 USDT

This initiative is designed for users engaging in real trading activity.

• Start: March 30, 2026, 00:00 (UTC+8)

• End: April 13, 2026, 00:00 (UTC+8)

• Duration: 14 days

As AI agents become more capable, their effectiveness must be tested beyond controlled environments.

GLM-5 Trading Odyssey provides a structured environment for real market validation.

Participants can begin by building and deploying their first agent.

VergeX is an AI trading harness layer that turns human intent into AI generated strategies, executed safely by AI agents in real markets.

Z.ai, founded in 2019, is a leading foundation model company behind GLM model series. GLM-5 ranks #1 on the Artificial Analysis Intelligence Index among all open-source models while GLM-5.1, soon to be open-sourced, delivers a 50% performance boost over its predecessor.

AI has made strategy generation easier than ever.

With large models, coding copilots, and AI agents, traders can now generate strategies faster than at any point in history. What once required weeks of quantitative research can now be done in minutes.

But strategy generation is not the same as trading.

Most AI trading experiments remain at the research stage. Strategies live in notebooks, backtests, or demo environments, but rarely make it into stable production systems.

The reality is simple:

The hardest part of trading is not generating strategies,
but running them reliably in real markets.

Production trading requires a full stack of infrastructure: reliable execution, risk controls, monitoring systems, and deterministic behavior.

Without these layers, strategies rarely make it beyond research environments.

As AI dramatically accelerates strategy creation, the real bottleneck is shifting elsewhere.

Traditional trading workflows typically look like this:

Research
↓
Strategy Design
↓
Manual Deployment
↓
Execution

This model relies heavily on human operators and does not scale efficiently.

In the age of AI, the trading stack is evolving:

AI Strategy Generation
↓
Validation & Risk Controls
↓
Production Execution
↓
Monitoring & Iteration

However, most AI trading tools today focus only on the first step: generating strategies.

The harder parts — validation, execution infrastructure, monitoring, and lifecycle management — are often overlooked.

Yet these are exactly the layers that determine whether a strategy can survive in production.

Trading systems are undergoing three structural shifts.

First, strategy generation is becoming commoditized.

AI tools are significantly lowering the barrier to creating trading strategies.

Second, execution complexity is increasing.

Liquidity is fragmented across exchanges and asset classes, making reliable execution infrastructure more critical than ever.

Third, trading systems are becoming autonomous.

Instead of relying on human operators, the next generation of trading systems will run continuously as automated systems.

In this environment, the core of trading is no longer just the strategy itself, but the system that runs it.

AI trading is increasingly an infrastructure problem.

VergeX is an AI-native trading operating system and harness layer designed for production environments.

Unlike tools that focus only on strategy generation, VergeX provides a complete trading harness that takes strategies from research to real-world deployment.

The system is built around three core layers.

Strategy Layer

AI-assisted strategy design, including natural language strategy generation and automated strategy workflows.

Validation & Risk Layer

Backtesting, simulation, and risk constraints — forming the control layer of the harness before strategies enter live markets.

Execution Layer

Deterministic execution, exchange connectivity, and real-time monitoring — ensuring strategies run reliably within the production harness.

Together, these layers form a unified AI trading harness for building, validating, and running strategies in real markets.

VergeX bridges the gap between strategy generation and production trading by providing the missing harness layer.

VergeX builds upon the research foundation of the NoFx ecosystem.

NoFx began as an open-source research initiative exploring AI-driven trading and multi-agent strategy experimentation. The project attracted a global community of contributors and has reached over 10,000 stars on GitHub.

Through this process, one insight became clear:

Open research accelerates experimentation, but production trading requires a robust harness to constrain and run strategies safely.

NoFx explored the research layer.
VergeX builds the harness for production.

Production systems require stability, risk control, and continuous iteration.

Through building NoFx, we realized that research is easy to scale — production is not.

For this reason, VergeX is launching in an invite-only beta.

This allows us to work closely with early builders and traders to refine the harness in real-world conditions.

Our goal is building a reliable foundation for AI-native trading.

Several trends are converging.

AI agents are becoming capable of generating strategies and interacting with complex systems. At the same time, global financial markets are increasingly accessible through APIs and open trading platforms.

These developments are enabling a new generation of trading infrastructure.

Instead of isolated strategies managed manually, trading systems are evolving into autonomous systems that continuously generate, evaluate, and deploy strategies.

In this environment, the harness layer becomes critical.

Over the next five years, trading infrastructure may undergo a fundamental transformation.

Autonomous Strategy Agents

AI agents running inside a shared harness, continuously generating, evaluating, and deploying strategies across global markets.

Strategy Ecosystems

Builders and developers contributing strategies that run on a common trading harness.

AI Native Financial Systems

Trading platforms designed around harnessed autonomous systems rather than manual operation.

VergeX is being built with this future in mind.

Apply for Beta
https://vergex.trade/

X
https://x.com/vergex_ai

Discord
https://discord.com/invite/vergex

Telegram
https://t.me/VergeX_EN

Trading digital assets involves risk. Only participate if you understand the risks and are comfortable with potential losses.