Pinned evaluation regime. Public participation. Deterministic scoring.

C.C.O -Cheaper FasterGPU compute

Live data, guide, docs, and links for CCO.

Open live data Read the guide

Network context

What CCO targets

GPU compute, LLM scale, measured gains

Show it visually.

Close-up photo of an integrated circuit on a microchip

GPU

Pinned hardware

Pinned hardware keeps runs honest.

Diagram of a multi-headed attention mechanism

LLM

Attention path

Attention drives the roadmap.

Butterfly diagram of a fast Fourier transform

FFT

Spectral route

Transforms push past dense matmul.

Visual sources: chip photo by Jon Sullivan / PD Photo.org (public domain); attention diagram by dvgodoy (CC BY 4.0); FFT butterfly diagram by Yangwenbo99 (CC BY-SA 4.0).

Participants

For miners, contributors, and researchers

Public participation across the CCO workflow.

For miners

Submit measured improvements

Open a PR with a scorecard.

Read the contribution flow

For contributors

Fixes and strategy work both matter

Fix and feat lanes differ.

Open CONTRIBUTING.md

For researchers

Track the operator class, not just one benchmark

Track matmul to attention.

Read the whitepaper

How CCO works

Submit a strategy. Measure it honestly.

Matmul now. Attention later.

Read the rule

All cost axes must drop while accuracy holds.

Build or fix

Add, tune, or fix.

Run the scorer

Produce one scorecard.

Publish and verify

Queue, replay, verify.

Live reference state

Queue, frontier, and recent results

Public queue and sealed results.

Queue updated

Results updated

Queued PRs

...

Tracks live

...

Pinned GPU

Published evals

...

No queued PRs right now.

No frontier data available.

No published evaluation results yet.

View the full live dashboard

Docs, landing pages, and repository surfaces

Core docs, public pages, and repository surfaces.

Whitepaper

Vision and roadmap.

Open resource

Contributing guide

PR lanes and checks.

Open resource

Benchmarks

How results are made.

Open resource

Eval README

Scorer and queue flow.

Open resource

Spectral token mixing

Why spectral matters.

Open resource

Spectral vs. matmul explained

Tradeoffs and context.

Open resource

CCO intro

Public overview.

Open resource

GitHub repository

Code, docs, eval.

Open resource

Gittensor

SN74 context.

Open resource

C.C.O -Cheaper FasterGPU compute

GPU compute, LLM scale, measured gains

Pinned hardware

Attention path

Spectral route

For miners, contributors, and researchers

Submit measured improvements

Fixes and strategy work both matter

Track the operator class, not just one benchmark

Submit a strategy. Measure it honestly.

Read the rule

Build or fix

Run the scorer

Publish and verify

Queue, frontier, and recent results

Queue feed

Results feed

Queue preview

Frontier

Recent published results

Docs, landing pages, and repository surfaces

Core docs

Whitepaper

Contributing guide

Benchmarks

Technical notes

Eval README

Spectral token mixing

Spectral vs. matmul explained

Related pages

CCO intro

GitHub repository

Gittensor