Cortex (SN100) Gives Bittensor Subnets a Consumer-Facing Front Door With Chat, Code, and Bot

Bittensor has no shortage of subnets building AI models, inference systems, agents, and other AI infrastructure.

The problem has always been getting all of that technology in front of ordinary users in a product they can use.

Cortex (SN10) is taking on that role.

Rather than asking users to navigate individual subnets and figure out how to use what they produce, Cortex brings different pieces of the Bittensor ecosystem together and packages them into polished, consumer-facing applications.

Cortex Chat launches with the Cortex 1 Mini model, followed by Cortex Code for developers and Cortex Bot for persistent AI agents.

The result is a single consumer-facing brand through which users can access AI products built on top of Bittensor’s broader subnet ecosystem.

Pricing is also positioned directly against ChatGPT and Claude, with a Free plan at $0 and premium options extending to a custom Pro tier.

Three Products, One Consumer Interface

Cortex currently packages three applications under a single account so that users do not need to interact with separate products and subnet interfaces.

1. Cortex Chat serves as the primary conversational layer: Talk to models, write with them, and generate images through one interface that hosts the Cortex model family.

2. Cortex Code operates directly inside your repository: No pasted diffs, no chat-with-a-different-icon, just an agent that opens sessions in isolated environments per repository.

3. Cortex Bot creates persistent agents with their own Linux computers: Name a bot, give it a look, assign a job, and it keeps working across sessions on its own machine.

Across these three products, Cortex gives Bittensor a common interface through which users can consume it.

How the Whole Stack Works

The Cortex architecture routes work through four connected layers, each with a distinct role in delivering the final user experience.

1. SN100 improves the models continuously through competitive mining: Every training cycle strengthens the same lineage rather than producing disconnected checkpoints.

2. Cortex applications sit on top as the user-facing surface: Chat, Code, and Bot each solve a distinct problem rather than being the same tool with different icons.

3. The Cortex Foundation coordinates the whole pipeline end to end: Model training, adaptation, product deployment, and ecosystem coordination all sit under one entity.

Every product uses the same Cortex account, with a single sign-in procedure to access Chat, Code, and Bot.

Pricing Is Level With ChatGPT and Claude

Cortex published a four-tier structure that mirrors every centralised alternative on the market at consumer price points.

Plan	Cortex	ChatGPT	Claude
Free	$0/month, Cortex 1 Mini, web search, code runs	$0/month, GPT-5.6 Terra with limits	$0/month, limited daily usage
Entry	Go: $8/month, more messages, priority	Go: $8/month, more messages, ads	No equivalent tier
Standard	Plus: $20/month, projects, saved work	Plus: $20/month, GPT-5.6 Sol flagship	Pro: $20/month, all models
Premium	Pro: custom, unlimited, SSO, admin	Pro: $100/month, usage multiples	Max: $100/month, usage multiples

Giving Bittensor a Front Door

For years, Bittensor’s biggest strength, its large collection of specialised subnets building different pieces of the AI stack, has also created a usability problem.

There are many things being built, but they are spread across individual subnets, dashboards, APIs, and technical interfaces.

Cortex provides a different approach: bring the ecosystem’s technology together and put it behind consumer applications.

Chat, Code, and Bot are therefore more than three standalone products. They represent a potential distribution layer for the wider Bittensor ecosystem, giving subnet builders a path from building AI infrastructure to having that technology reach actual users.

If Cortex succeeds at that, its most important contribution to Bittensor may not be another AI model. It may be becoming the front door through which ordinary users finally experience what the network’s many subnets are building.

➛ Access Cortex Here.