Ridges: How an Open Tournament Beat Billion-Dollar AI Labs
Shakeel set out with the audacious idea of running an open, incentivized tournament to build the best software engineering agent. Four months later, that experiment has produced the top-ranked fully open-source software engineering agent. The agent has even outperformed Anthropic’s Claude 4 with tool support, showing that a small crypto-AI project can stand toe to toe with billion-dollar labs.
So how did Shakeel and three undergrad interns, with no outside capital or prior AI experience, pull this off? With a system they built called Ridges, where open competition and incentives did all the heavy lifting.
Ridges works like a simple SWE leaderboard. Developers submit an agent harness and scaffolding as a single Python file, and the agent must be built on an open-source model. Every submission is also open-sourced and benchmarked on SWE-bench, the standard dataset for measuring real-world software engineering capability by resolving issues from active GitHub projects.
The highest-scoring agent each day earns the prize pool, winner-take-all style, which is currently $55,000 per day. Rewards stream out every minute, with the leader continuing to collect until someone dethrones them from the top by submitting a better SWE agent.
Where Does the Prize Pool Come From?
Ridges runs on Bittensor, an open platform for building AI projects. Every project on Bittensor—called a subnet—issues its own token, which is used to economically incentivize and reward contributors. This setup is known as a crypto-economic system, where funding and ownership are baked directly into the protocol.
In Ridges, most of the tokens emitted each day go directly to the top-performing agent developer. They can sell those tokens or hold them as a long-term stake. Because the token is liquid from day one, its market price determines the value of the rewards.
Many contributors choose to hold, and investors are drawn in, because owning the token means owning a piece of the system. You can think of it like if Linux or another open-source project had a built-in ownership layer. Subnet founders decide what rights come with holding their token, but the most common are governance, economic participation, and profit sharing.
How a Small Crypto-AI Project Beat Well-Funded Incumbents
Ridges’ real edge comes from ownership and openness. In most open-source projects, contributors come and go. On Ridges, every developer who submits an agent also earns a stake in the system itself. That ownership works like equity in a company, which makes people go further, fix problems, and help improve the platform because they feel invested in its success.
Another driver is access to global talent. Ridges is agnostic to credentials, and only cares about results. In the last month alone, more than 800 people submitted over 5,600 agents, from solo developers to researchers inside AI labs, all competing on the same leaderboard. And because every submission must be open-sourced, each new agent becomes raw material for the next. Ideas stack, iteration speeds up, and progress compounds.
Finally, the tournament never ends. Token rewards flow continuously, keeping the best agent devs in the world locked in an ongoing race to build the strongest agent and adapt instantly to new models and methods as they appear.
This flywheel of ownership, open competition, and a never-ending tournament is Ridges’ superpower. It’s what pushed the system to the frontier of performance, faster than billion-dollar labs.
A New Model for Developing Open Source
Open source has always struggled with two things. Funding is scarce, with most projects relying on volunteers or a few sponsors. And alignment is fragile, since forking often pulls effort in different directions and weakens the sense of a shared mission.
Ridges shows how crypto-economic systems can solve both. Tokens provide built-in funding, and ownership keeps contributors working toward the same goal. Instead of splintering, the work stacks, each improvement adding to the next. The more capital that flows in, the bigger the prize pool becomes, and the stronger the pull is to keep building together.
Zooming out, what Ridges has actually created is a new kind of open-source engine: one with its own fuel, its own alignment, and the power to sustain progress at scale.
The Road Ahead
With Ridges now leading across open AI and narrowing the gap with closed systems, the team is moving towards its next phase of development: building a product powered by the underlying agents on the subnet. The first iteration will be a terminal-based system, similar to Claude Code or OpenAI’s Codex.
Because Ridges’ agents are built on open models, they can deliver comparable performance at a fraction of the cost—roughly 90 percent cheaper than incumbent offerings. Running all 500 SWE-bench problems with the top Ridges agent costs just $1.26, compared to $94 for Claude Opus.
The team is following a well-tested playbook in commercial open source. The core coding agent will remain open and community-driven, while premium features and managed cloud services will provide the monetization layer. Just as Red Hat, MongoDB, and GitLab turned open foundations into scaled businesses, Ridges aims to do the same for AI coding agents.
Shakeel has already proven he can execute at breakneck speed. In just four months he has grown a contributor base of nearly 1,000 developers and delivered the leading software engineering agent. For anyone who has dismissed crypto-AI, Ridges is proof to the contrary, and in our opinion, the most impressive achievement in the space so far.
This content is provided for informational purposes only and does not constitute investment advice or a recommendation to buy or sell any security. Unsupervised Capital holds a position in TAO and may hold positions in the subnet tokens or other digital assets discussed herein and may buy, sell, or change positions at any time. Past performance is not indicative of future results. Digital assets involve substantial risk, including potential total loss of capital. Consult your own advisers regarding any investment decisions.
Investment in digital assets and blockchain technology involves substantial risk and may result in partial or total loss of investment. Digital assets are subject to extreme price volatility, regulatory uncertainty, and market manipulation. Past performance is not indicative of future results. This memorandum is for informational purposes only and does not constitute investment advice, an offer to buy or sell securities, or a solicitation of any offer to buy or sell securities. Recipients should consult their own advisors before making investment decisions. This memorandum contains projections and estimates based on current expectations. Actual results may differ materially from those projected. All forward-looking statements are subject to risks and uncertainties. Unsupervised Capital Management, LLC makes no representations regarding the accuracy or completeness of information herein. Third-party information has not been independently verified. To the maximum extent permitted by law, Unsupervised Capital Management, LLC shall not be liable for any damages arising from this memorandum. By reviewing this memorandum, you acknowledge these disclosures. This information is subject to change without notice.
