Neuron registration trends
The alpha price in τ, day by day, and what moved it each way.
Log change over the last 30 days −2.20%: protocol buys +7.59%, root sells −1.57%, users net −8.02%, cross-subnet −0.20%, unattributed +0.0002%.
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.
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).
2026-11-13 UTC, as of ≈02:19 UTC (block 9,494,383)
The day isn’t over yet, so these figures keep changing. They are in no daily total, average or rank.
Holder & circulating-alpha metrics as of (daily close)
Autonomous drone pilots are developed behind closed doors with proprietary test environments, making it impossible to compare solutions or prove which is best. The same hardware and software improvements get rediscovered repeatedly, and real-world deployment requires pilots trained on diverse, realistic scenarios — not static test sets. A decentralized benchmark platform where anyone submits a pilot, flies against fresh procedurally generated worlds every epoch, and scores against a public leaderboard solves both problems: results are transparent and reproducible, and the diversity of competitor approaches drives innovation across the field.
Swarm Subnet 124 implements this as a network where miners train drone navigation policies locally in a standard gym environment, package them as sandboxed Docker containers, and submit them for evaluation. Validators run the same simulation engine, the same procedurally generated seeds, and the same scoring rules — ensuring every submission is judged identically.
New worlds are generated every 14 days (per epoch), seeded identically across all validators, so no pilot can memorize and no private test set gives unfair advantage. Results feed a public leaderboard and reward structure (King of the Hill) where the strongest pilots earn continuously, and breakthroughs that move the frontier keep earning even after being surpassed.
The subnet covers three distinct missions: solar park patrol with thermal and visual detection (85% of rewards), multi-drone search and rescue coordination (10%), and single-drone victim hover confirmation (5%). Each family maintains its own champion lineage and scoring rules, and validators enforce that every miner and every benchmark run use identical code, seeds, and physics simulation — achieved through pinned dependency versions, deterministic world generation seeded from the map seed, and local re-computation of weights by validators from published champion scores.
Auto-generated from the GitHub repository and may not reflect the latest changes.
Swarm Subnet 124 is an open autonomous drone navigation benchmark and competition platform on Bittensor where miners train AI pilots to fly drones across procedurally generated environments, and validators score submissions across three active challenge families: Swarm Sentinel (85% emissions, solar park patrol with thermal/visual detection), Swarm Search and Rescue (10% emissions, multi-drone victim location), and Search and Rescue (5% emissions, single-drone victim hover confirmation).
Miners train policies in a local gym environment (swarm-drone-gym), package them as Docker-sandboxed Python RPC servers, and submit to the Swarm backend. Validators run the benchmark engine (PyBullet with swarm-bullet3), which loads procedurally generated seeds, spawns Docker containers with miner models, feeds observations (depth/RGB/state) at fixed rates (10 Hz for Sentinel, 50 Hz control/decision split for others), collects actions, simulates physics (50 Hz), and scores per family rules. Both miners and validators use the same simulator, same seeds (per epoch), and same scoring logic for reproducibility.
The backend holds champion archives and metadata; validators recompute King of the Hill weights locally from per-family champion scores and the published formula, with no trusted third party.
Each of three active families (Swarm Sentinel 85%, Swarm SAR 10%, Search and Rescue 5%) runs its own King of the Hill lineage. A model is crowned by passing the full benchmark (1,000 or 2,000 seeds depending on family) and clearing the crowning floor. Rank weighting (0.7^rank for position in window) rewards fresh wins; log-headroom gain bonus (up to +30% within rank) rewards moving the frontier.
Gain is locked at crowning; rank shifts as new kings arrive. The formula ensures the reigning champion always holds the largest seat, and past kings age out after 5 (or 10 in Sentinel) dethronings. If a family's window has no eligible king with positive gain, that family's slice burns.
Epochs run 14 days; the reigning champion is re-scored on new epoch seeds but keeps the crown unless dethroned on the same epoch.
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)
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.
Not yet verified: Notes
Not yet verified: 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 2 Nov 2026
Install Swarm CLI and dependencies in a virtual environment
uv venv && source .venv/bin/activate && uv pip install -e .
Check that your environment is ready for benchmarking (Python 3.11+, Docker, required Python modules)
swarm doctor
Package your drone model as a submission ZIP for your chosen family
swarm model package --source ./my_model --family-id cf_search_and_rescue
Verify submission ZIP against Swarm rules before benchmarking
swarm model verify --model Submission/submission.zip
Score your model locally against 1,000 seeds to test before network submission
swarm benchmark --model Submission/submission.zip --family-id cf_search_and_rescue --workers 4
Submit your model to the network from your hotkey (commits hash on-chain, uploads archive)
swarm model submit --source ./my_model --family-id cf_search_and_rescue --wallet.name my_cold --wallet.hotkey my_hot
Install Swarm validator and dependencies in a virtual environment
uv venv && source .venv/bin/activate && uv pip install -e .
Check environment is ready: Python 3.11+, Docker daemon running, sandbox lockdown binaries (nsenter, iptables) installed and with correct permissions
swarm doctor
Start the validator process; it will initialize Docker evaluator, load state, and begin scoring submitted models
python neurons/validator.py
Optional: run the local dashboard to monitor validator runtime, events, and scoring in real time(optional)
swarm monitor
| Day (UTC) | Share |
|---|---|
| 2026-11-12 | 0.64% |
| 2026-11-11 | 0.61% |
| 2026-11-10 | 0.60% |
| 2026-11-09 | 0.60% |
| 2026-11-08 | 0.60% |
| 2026-11-07 | 0.61% |
| 2026-11-06 | 0.63% |
| 2026-11-05 | 0.64% |
| 2026-11-04 | 0.67% |
| 2026-11-03 | 0.68% |
| 2026-11-02 | 0.67% |
| 2026-11-01 | 0.68% |
| 2026-10-31 | 0.70% |
| 2026-10-30 | 0.68% |
| 2026-10-29 | 0.68% |
| 2026-10-28 | 0.69% |
| 2026-10-27 | 0.69% |
| 2026-10-26 | 0.72% |
| 2026-10-25 | 0.72% |
| 2026-10-24 | 0.71% |
| 2026-10-23 | 0.77% |
| 2026-10-22 | 0.81% |
| 2026-10-21 | 0.76% |
| 2026-10-20 | 0.75% |
| 2026-10-19 | 0.71% |
| 2026-10-18 | 0.72% |
| 2026-10-17 | 0.73% |
| 2026-10-16 | 0.76% |
| 2026-10-15 | 0.75% |
| 2026-10-14 | 0.78% |
| 2026-10-13 | 0.75% |
| 2026-10-12 | 0.73% |
| 2026-10-11 | 0.75% |
| 2026-10-10 | 0.78% |
| 2026-10-09 | 0.80% |
| 2026-10-08 | 0.82% |
| 2026-10-07 | 0.85% |
| 2026-10-06 | 0.85% |
| 2026-10-05 | 0.83% |
| 2026-10-04 | 0.77% |
| 2026-10-03 | 0.76% |
| 2026-10-02 | 0.81% |
| 2026-10-01 | 0.78% |
| 2026-09-30 | 0.88% |
| 2026-09-29 | 0.82% |
| 2026-09-28 | 0.80% |
| 2026-09-27 | 0.82% |
| 2026-09-26 | 0.94% |
| 2026-09-25 | 0.88% |
| 2026-09-24 | 0.87% |
| 2026-09-23 | 0.90% |
| 2026-09-22 | 0.94% |
| 2026-09-21 | 0.90% |
| 2026-09-20 | 0.94% |
| 2026-09-19 | 0.95% |
| 2026-09-18 | 1.00% |
| 2026-09-17 | 1.02% |
| 2026-09-16 | 1.08% |
| 2026-09-15 | 1.09% |
| 2026-09-14 | 1.13% |
| 2026-09-13 | 1.12% |
| 2026-09-12 | 1.14% |
| 2026-09-11 | 1.12% |
| 2026-09-10 | 1.10% |
| 2026-09-09 | 1.16% |
| 2026-09-08 | 1.30% |
| 2026-09-07 | 1.32% |
| 2026-09-06 | 1.37% |
| 2026-09-05 | 1.43% |
| 2026-09-04 | 1.42% |
| 2026-09-03 | 1.46% |
| 2026-09-02 | 1.37% |
| 2026-09-01 | 1.18% |
| 2026-08-31 | 1.24% |
| 2026-08-30 | 1.21% |
| 2026-08-29 | 1.23% |
| 2026-08-28 | 1.31% |
| 2026-08-27 | 1.29% |
| 2026-08-26 | 1.26% |
| 2026-08-25 | 1.30% |
| 2026-08-24 | 1.30% |
| 2026-08-23 | 1.24% |
| 2026-08-22 | 1.28% |
| 2026-08-21 | 1.30% |
| 2026-08-20 | 1.29% |
| 2026-08-19 | 1.32% |
| 2026-08-18 | 1.24% |
| 2026-08-17 | 1.22% |
| 2026-08-16 | 1.21% |
| 2026-08-15 | 1.18% |
