The chart follows the alpha price in τ. Each period starts at its opening price and steps through the mechanisms below, in a fixed order, to its closing price: a step that rises pushed the price up, one that falls pushed it down, and its height is how far it moved the price. A thin line carries each close into the next period’s open; it is dashed where the two differ. The faint candle behind each period is the market’s own open, high, low and close, for context.

With many periods on screen (90 days, or hours), the steps would be too thin to read, so the chart draws the price as a line instead and colours each period by the mechanism that moved it most. Hover a period for the full step-by-step breakdown.

Protocol buy: The chain spending excess emission TAO to buy alpha from the pool. A permanent protocol buyback; it pushes the price up.
Root sell: Root-dividend claims sold into the pool; they push the price down.
User buy / sell: Staking into and unstaking from this subnet.
Cross-subnet: Swaps from other subnets routed through this pool.
Residual: What the decomposition cannot attribute to a mechanism. It is always shown, never folded into another part.

Not here, because they do not move the spot price: the per-block TAO injection into the pool, and burns.

All figures are log changes, not simple percentage changes, so the parts add up exactly to the total; on big moves the two differ (a 10% log fall is a 9.5% price fall).

Today so far

2026-11-07 UTC, as of ≈11:07 UTC (block 9,453,830)

Price, today so far: 0.004464 τ
Volume, today so far: 162.78 τ
Market cap incl. burned, today so far: 5.37K τ
TAO in pool, today so far: 3.10K τ
Emission share, today so far: —not available yet

The day isn’t over yet, so these figures keep changing. They are in no daily total, average or rank.

Subnet Configuration

Pool & Mechanism

AMM Pool

Balancer

Mechanism

dTAO

Access

Registration

Open

Neurons

256/256Competitive

Validators

5/6459 open

Economics

Owner Cut

18.00%

Weights

Disabled

Disabled

1–65535

Timing

Tempo

360~1.2h

Immunity Period

5,000~16.7h

Activity Cutoff

5,000~16.7h

Market & Economics

Market

Market Cap

3,834.8658 τ

FDV

8,464.3827 τ

TVL

6,190.1524 τ

Lifetime Volume

260.53K τ

Holders

1,390

Owner Locked

2,002.3405 ვ(0.23% of circulating)Perpetual

14d price

Revenue / Recycle

2.46×

last day · Net positive

Emissions

0.04%of network emissions

TAO emitted to this subnet

Sent into its pool each day: its share of network emissions

1.4613 τ/day

1.4613 τ per day

Daily distribution

Alpha paid out each day, valued at the pool price

7,200.0000 ვ/day

≈ 32.1406 τ·32.1406 τ per day

Miners49.47%

3,561.7286 ვ ≈ 15.8995 τ/day

Validators32.53%

2,342.3044 ვ ≈ 10.4560 τ/day

Owner18.00%

1,295.9670 ვ ≈ 5.7852 τ/day

Stake 1,000 τ here

756.7841 ვ/day≈ 3.3783 τ·3.3783 τ per day

From the validator share, before commission

121.46%staker APY

· 30d avg 121.46% · Stable

90d

Pool emits 1.0000 ვ/block at 0.004464 τ/ვ. Yield varies with stake weight, pool price, and emission rate.

For Miners

49.47% of emission

Slots

256/256Competitive

Burn Cost

0.0055 τ

Est. per Miner

29.6811 ვ/epoch≈ 0.1325 τ

Pool / Epoch

178.0864 ვ≈ 0.7950 τ

For Validators & Stakers

32.53% of emission

Validator Slots

5/6459 open

Est. per Validator

23.4230 ვ/epoch≈ 0.1046 τ

Registration Activity

Neuron registration trends

This Interval

24h: 50 registrations (50 burn · 0 PoW)

Target / Interval

Slots Available

Competitive

256/256

Neurons

256

Development Activity

Connito-AI/ConnitoPython

5134 contributors119 commits/30dLast commit 7d ago

MIT

Top Contributors

Recent Releases

v0.1.31189d ago v0.1.1197d ago v0.0.6199d ago v0.0.5200d ago v0.0.4201d ago

What ConnitoAI does

Building high-performance language models typically requires centralizing training infrastructure and manually coordinating updates across versions. This creates bottlenecks in iteration speed, siloes expertise across teams, and makes it difficult to assess which improvements genuinely generalize versus overfit to narrow benchmarks. A decentralized network of independent trainers competing under measurable, verifiable incentives can explore different training strategies in parallel while a consensus mechanism filters for genuine improvements.

SN102 coordinates this through a chain-driven cycle structure where miners train small residual expert networks on top of a frozen base model, competing on task-specific loss improvement while validators measure generalization loss on held-out replay data. Each validator independently scores submissions using identical eval procedures seeded from chain state, preventing collusion. Winners are published to shared storage and become the starting point for the next round, creating a collective ratchet of improvements.

The mechanism uses Bittensor's chain coordination to synchronize phase boundaries (distribute, train, commit, validate, merge) across all participants, avoiding the need for a central scheduler. Validators maintain rolling score histories and emit chain weights proportional to long-term performance, ensuring consistent performers receive sustained allocation while one-off flukes do not.

Auto-generated from the GitHub repository and may not reflect the latest changes.

About

active

Connito SN102 is a subnet that trains residual expert networks on top of a frozen base language model (Qwen 3.6-35B). Miners train small mixture-of-experts (MoE) modules in parallel cycles, competing on task performance and generalization. Validators score submissions on task loss improvement and c4 replay loss isolation, publishing winners that subsequent miners build upon.

The subnet coordinates through chain-based phases and maintains a consensus-driven weight emission system via stake-weighted aggregation.

How it works

The subnet operates on fixed-length cycles (blocks measured in chain time) with nine sequential phases: Distribute (miners download round winner), Train (miners train residuals), MinerCommit1/2 (miners sign and upload), Submission (validators score all miners), Validate (scores finalized), Merge (validators allreduce gradients via DHT), ValidatorCommit1/2 (validators publish winner and weights). Miners train small residual expert MLPs that sit beside every MoE block of the frozen Qwen 3.6-35B base model. Validators run on Docker with auto-update via Watchtower.

Chain commits from miners are validated (signature, hash, residual spec) before scoring. Winners are published to HuggingFace by each validator and downloaded via chain-pinned coordinates in the next Distribute. Consensus is reached via stake-weighted chain-weight aggregation across all validators.

Incentive mechanism

Each cycle, validators measure every miner's residual on two eval splits: (1) task data (pmc) for task gain, (2) c4 replay data for generalization. Raw score = task_gain - replay_penalty - |task_usage - 0.2| - replay_usage, where penalties are subtracted, and final score = raw - base_raw (so the pretrained base scores exactly 0). If c4 loss rises >0.02 above the base, the miner is stopped (score = -inf, not ranked).

Top-3 miners by score receive rank-based rewards; miners 4+ get 0. Hotkey rotation resets history. Weight Group 1 (73.5% on wire) requires ≥3 samples, ≥3 distinct rounds in last 5 cycles, and score_avg > 0.

Weight Group 2 (1.5% on wire) requires ≥1 sample and score_avg > 0. If no miner qualifies for G1, the 73.5% redirects to UID 0 (subnet owner), not burned.

The sections above were auto-generated from the GitHub repository and may not reflect the latest changes. View repository (github.com, opens in a new tab, external site)

On-Chain Identitysigned on-chain by subnet owner

ვ

ConnitoAI

Vin · Georgian letter · pronounced “vin”

Contributors train specialized expert modules that are aggregated into powerful AI systems, without massive centralized compute.

Owner: 5Hf5D…JZYA
Hotkey: 5EEin…KkC9
Owner Cut: 18.00%
Contact: [email protected]

Registered: 7 months ago(Mar 28, 2026)
Block: 7,840,965
Slot history: Every subnet in SN102
Neurons: 256/256
Tempo: 360 blocks

Website

URL: https://connito.ai/ (connito.ai, opens in a new tab, external site)
Status: Down403
TLS: Valid
Title: Just a moment...
Checked: 2026-11-07 04:02 UTC

Domain registered after this subnet started · Mar 31, 2026

Changes (10)

2026-11-07 04:02 UTCDown403Just a moment...
2026-11-06 04:01 UTCDown403Just a moment...
2026-11-05 04:02 UTCDown403Just a moment...
2026-11-04 04:01 UTCDown403Just a moment...
2026-11-03 04:03 UTCDown403Just a moment...
2026-11-02 04:02 UTCDown403Just a moment...
2026-11-01 04:02 UTCDown403Just a moment...
2026-10-31 04:02 UTCDown403Just a moment...
2026-10-30 04:02 UTCDown403Just a moment...
2026-10-29 04:02 UTCDown403Just a moment...

Subnet History

SubnetIdentitySet 2026-08-04 17:19 UTC

Block 8,772,278Extrinsic

SubnetIdentitySet 2026-05-13 13:18 UTC

Block 8,175,009Extrinsic

Emissions started 2026-04-03 17:46 UTC

Block 7,888,925Extrinsic

Subnet registered 2026-03-28 01:14 UTC

Block 7,840,965Extrinsic

SubnetIdentitySet 2026-03-28 01:14 UTC

Block 7,840,965Extrinsic

Recent Activity

SubnetIdentitySet Block 8,772,278 Extrinsic2026-08-04 17:19 UTC
SubnetIdentitySet Block 8,175,009 Extrinsic2026-05-13 13:18 UTC
FirstEmissionBlockNumberSet Block 7,888,925 Extrinsic2026-04-03 17:46 UTC
NetworkAdded Block 7,840,965 Extrinsic2026-03-28 01:14 UTC
SubnetIdentitySet Block 7,840,965 Extrinsic2026-03-28 01:14 UTC

This information was auto-extracted from the GitHub repository and may be incomplete or out of date. Always refer to the subnet's GitHub repository (github.com, opens in a new tab, external site) for the latest instructions.

Hardware Requirements

Miner

GPURequired

VRAM48 GB

RAM62 GB

Storage100 GB

OSLinux with NVIDIA driver ≥ 570

Not yet verified: GPU model, GPU count, Network, Notes

Validator

GPURequired

GPU ModelRTX A6000 or L40S, 48 GB minimum

VRAM48 GB

RAM66.57 GB

Storage100 GB

OSLinux with NVIDIA driver ≥ 570

Not yet verified: GPU count, Network, Notes

Machine-extracted from the subnet's repository docs; may be wrong or out of date. Check the repository before buying or renting hardware. Open the repository (github.com, opens in a new tab, external site) · Extracted 19 Oct 2026

Miner Setup

Create Python 3.10 virtual environment with headers for Triton compilation

uv venv --seed --managed-python -p 3.10 .venv && source .venv/bin/activate

Install Python dependencies including torch 2.10.0 (cu128), transformers 5.18.0, flash-linear-attention 0.5.2

pip install -r requirements.txt --extra-index-url https://download.pytorch.org/whl/cu128

Pre-download the 67 GiB base model to avoid first-round download delays(optional)

hf download Qwen/Qwen3.6-35B-A3B --revision 995ad96eacd98c81ed38be0c5b274b04031597b0

Generate validator.yaml config file with wallet credentials

python connito/shared/config.py create_config --role miner --coldkey_name <name> --hotkey_name <name>

Set environment variables for HuggingFace token (write access) and bittensor wallet path

export HF_TOKEN='hf_...' && export BITTENSOR_WALLET_PATH=~/.bittensor/wallets

Verify subnet 102 exists and register your hotkey via proof-of-work

btcli subnet list && btcli subnet pow_register --netuid 102 --wallet.name <coldkey> --wallet.hotkey <hotkey>

Start the miner training loop (trains and commits each cycle)

python -m connito.miner.train --path checkpoints/miner/<coldkey>/<hotkey>/foundation/config.yaml

Validator Setup

Install Docker Engine and Compose plugin

curl -fsSL https://get.docker.com | sudo sh && sudo usermod -aG docker $USER && newgrp docker

Install NVIDIA Container Toolkit for GPU access in containers

distribution=$(. /etc/os-release; echo $ID$VERSION_ID) && curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg && curl -fsSL https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list | sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list && sudo apt-get update && sudo apt-get install -y nvidia-container-toolkit && sudo nvidia-ctk runtime configure --runtime=docker && sudo systemctl restart docker

Authenticate to GitHub Container Registry (only if image is private)(optional)

echo '<github_pat>' | docker login ghcr.io -u <username> --password-stdin

Generate validator.yaml and Docker .env file with wallet credentials

python connito/shared/config.py create_config --role validator --coldkey_name <name> --hotkey_name <name>

Pre-download the 67 GiB base model to the shared cache(optional)

HF_HOME=<DATA_DIR>/hf hf download Qwen/Qwen3.6-35B-A3B --revision 995ad96eacd98c81ed38be0c5b274b04031597b0

Start validator and Watchtower containers with auto-update enabled

cd connito/validator/docker && docker compose up -d

Verify validator boots and reaches first cycle phase (should see 'Commit new seed' within minutes)

docker compose logs -f validator | head -50

No hardware? Back this subnet by staking instead New to mining? Start with the setup guide Validator setup guide

Day (UTC)

2026-11-06

0.04%

2026-11-05

0.04%

2026-11-04

0.05%

2026-11-03

0.06%

2026-11-02

0.04%

2026-11-01

0.03%

2026-10-31

0.03%

2026-10-30

0.02%

2026-10-29

0.02%

2026-10-28

0.03%

2026-10-27

0.03%

2026-10-26

0.03%

2026-10-25

0.02%

2026-10-24

0.02%

2026-10-23

0.03%

2026-10-22

0.05%

2026-10-21

0.05%

2026-10-20

0.06%

2026-10-19

0.07%

2026-10-18

0.07%

2026-10-17

0.07%

2026-10-16

0.07%

2026-10-15

0.07%

2026-10-14

0.09%

2026-10-13

0.08%

2026-10-12

0.08%

2026-10-11

0.08%

2026-10-10

0.09%

2026-10-09

0.08%

2026-10-08

0.07%

2026-10-07

0.06%

2026-10-06

0.07%

2026-10-05

0.08%

2026-10-04

0.08%

2026-10-03

0.08%

2026-10-02

0.08%

2026-10-01

0.07%

2026-09-30

0.10%

2026-09-29

0.10%

2026-09-28

0.10%

2026-09-27

0.10%

2026-09-26

0.13%

2026-09-25

0.10%

2026-09-24

0.06%

2026-09-23

0.05%

2026-09-22

0.07%

2026-09-21

0.09%

2026-09-20

0.10%

2026-09-19

0.10%

2026-09-18

0.11%

2026-09-17

0.13%

2026-09-16

0.14%

2026-09-15

0.14%

2026-09-14

0.17%

2026-09-13

0.09%

2026-09-12

0.17%

2026-09-11

0.19%

2026-09-10

0.19%

2026-09-09

0.22%

2026-09-08

0.21%

2026-09-07

0.20%

2026-09-06

0.25%

2026-09-05

0.31%

2026-09-04

0.33%

2026-09-03

0.35%

2026-09-02

0.31%

2026-09-01

0.25%

2026-08-31

0.27%

2026-08-30

0.31%

2026-08-29

0.34%

2026-08-28

0.29%

2026-08-27

0.41%

2026-08-26

0.42%

2026-08-25

0.50%

2026-08-24

0.49%

2026-08-23

0.50%

2026-08-22

0.47%

2026-08-21

0.45%

2026-08-20

0.46%

2026-08-19

0.44%

2026-08-18

0.39%

2026-08-17

0.41%

2026-08-16

0.44%

2026-08-15

0.47%

2026-08-14

0.52%

2026-08-13

0.55%

2026-08-12

0.53%

2026-08-11

0.45%

2026-08-10

0.38%

2026-08-09

0.40%

distribution=$(. /etc/os-release; echo $ID$VERSION_ID) && curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg && curl -fsSL https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list | sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list && sudo apt-get update && sudo apt-get install -y nvidia-container-toolkit && sudo nvidia-ctk runtime configure --runtime=docker && sudo systemctl restart docker

What moved the price

Overview

Today so far

Subnet Configuration

Market & Economics

For Miners

For Validators & Stakers

Registration Activity

Development Activity

What ConnitoAI does

About

Website

Subnet History

Recent Activity

Participate

Hardware Requirements

Miner Setup

Validator Setup

What moved the price

Today so far

Subnet Configuration

Market & Economics

For Miners

For Validators & Stakers

What ConnitoAI does

About

Website

Subnet History

Hardware Requirements

Miner Setup

Validator Setup