Cortex Foundaτion

We just completed the largest decentralised LLM pre-training run in history: Covenant-72B. Permissionless, on Bittensor subnet 3. 72B parameters. ~1.1T tokens. Commodity internet. No centralized cluster. No whitelist. Anyone with GPUs could join or leave freely. 1/n

GLM-5 is LIVE on Chutes 🔥 @zai_org's new flagship. 744B params (40B active). Best-in-class open-source on reasoning, coding, and agentic tasks. 77.8% SWE-Bench Verified 73.3% SWE-Bench Multilingual 75.9% BrowseComp (w/ context manage) 50.4% HLE with tools 89.7% τ²-Bench Try it now → chutes.ai/app/chute/e51e…

AI coding tools are already a multi-billion dollar market. Cursor, Copilot, Claude Code… but almost everything is closed-source. We're launching Cortex on February 20th. Public beta, fully open source, continuously improved by Subnet 100 miners on Bittensor.

Tau Ninja is running on Basilica. basilica summon tau You need just two environment variables. A Chutes API key for inference and a @telegram bot token from @BotfatherTG. That's the entire setup for @const_reborn's self-modifying AI agent on decentralized compute. Under the hood, three subnets are each contributing their output to a single deployment. Basilica (SN39) provisions and runs the container. This is the compute surface where everything comes together. @chutes_ai (SN64) handles LLM inference. Every time Tau reasons about a task, plans its next action, or generates a response, that call routes through Chutes' decentralized inference network. @platform_tao (SN100) provides the baseagent framework. The agent scaffolding, tool execution, and state management that make Tau an autonomous agent rather than a chatbot. Because baseagent lives on a Bittensor subnet, miners are continuously incentivized to improve it. The agent framework itself gets better over time through market competition, and every agent built on it inherits those improvements. That's what makes this a Bittensor-native agent rather than just an agent deployed on Bittensor. Each subnet does one thing well. Basilica is the deployment layer that composes them into a working product. One command, seconds to deploy, and the result is a self-modifying AI agent running entirely on decentralized infrastructure. tau.ninja For support and setup help, join us on Discord: discord.gg/kUJFjTXRQx

For months, agent validation on our subnet relied on manual processes. Slow, limited, and hard to scale. Today, thanks to @chutes_ai, Subnet 100 has fully transitioned to autonomous agent validation. Faster, more reliable, and built to scale. This is just the beginning.

We’re now at 63.3%✅ on evaluated tasks. Next challenge: Bug Bounτy ! Miners report bugs & suggestions, earn points, and help improve the software. Our first CLI Agentic Coding drops next week.