Subnets / SN10
Pareton SN10
Pareton (subnet 10) is a Bittensor subnet held by 2,096 unique coldkeys. Its liquidity pool holds 13.5k τ, with the alpha token trading at 0.0043 τ. Roughly 0% of tracked holders are in profit — most bought in above today's price. Seller exhaustion is elevated: 85% of recent sellers have emptied their wallets, thinning the supply overhang. Full insider flow, whale tracking and the complete proprietary factor set for Pareton are available to SubnetStats subscribers.
ALPHA PRICE
0.0043 τ
IN TAO
HOLDERS
2,096
UNIQUE COLDKEYS
POOL DEPTH
13.5k τ
TAO IN POOL
7D MOMENTUM
-3.6%
ALPHA vs TAO
Project activity
@Muse Spark Open is now open for submissions on Pareton.
Focusing on the serving track of the Open Kernel Challenge — patch SGLang to serve Meta’s Muse Spark Open faster on 8× H200 GPUs at 16 concurrent requests.
Fixed 220-token outputs from agentic traces.
https://t.co/vkRukT
Excited to kick off a pilot with @LeadpoetAI.
Leadpoet will help us find AI companies dealing with rising GPU costs and inference bottlenecks, identify the right technical decision-makers, and run targeted outreach to turn those opportunities into customers.
SN71 🤝 SN10 https:/
Excited to be one of the new verified providers on the Liquid Inference network by @Architect_Fi
Inference is rapidly becoming commoditized and financialized, and Liquid Inference is well positioned to bring real-time price discovery and continuous competition to every token.
Get paid to make Kimi K3 faster on SGLang.
8× B300. 16 concurrent requests. Make inference faster while preserving model quality.
The campaign is live → https://t.co/IKQUkAZczf https://t.co/fwNY2de7Yf
We just purchased some RTX 5090s to expand our inference operations.
Some disassembly required. Stay tuned. https://t.co/hmjHesomWT
Cost per token keeps falling, and every year the total bill goes up anyway. Reasoning models think longer. Agents run in loops. Context windows keep growing. Demand for inference is growing faster than our efficiency at serving it.
The gap between those two curves is your margi
Specialist vs. Generalist in an AI World.
Specialists win once models commoditize. One model, one GPU class, and the margin shows up. Generalists win on the problems that sit between fields. Pure depth misses the bridge.
We sit in the middle. Narrow where the search has to comp
Join us now on Discord
https://t.co/B6MxYzkOC4
TheStreet just profiled three @Bittensor teams chasing the least crypto thing in AI: customers who pay.
Pareton makes AI faster and cheaper. @LeadpoetAI builds sales intelligence. @openroboto builds models for robots.
Full piece: https://t.co/aKpL5pmFkF
Also on Yahoo Finance,
The shovels are built. What's left is the right tool for the ground you're actually digging in.
We're building an open arena for that. Point it at your app, miners compete on the software that runs the model, and we only keep what makes that job faster and cheaper. Optimizing AI
Where we’re going:
• Inference optimization — for inference providers, agent companies, and anyone running AI at volume. Make workloads cheaper and/or faster. (SN53 partnership live, more in the works.)
• Exotic setup — optimize inference on non-standard hardware or model to unlo
Engy is getting faster.
Pareton now tunes how @engyai runs its models. Quicker replies. Same API on your side.
Engy is one of the fastest-growing inference networks on @bittensor and beyond. A big customer with serious traffic. That is the workload we were built for.
Come check out our talk on Tuesday 14:20 at the RAO Stage! @ExploitSummit
base:0xddde5965d49e69b6362bd8743e139a45dbfeb7b3 is live on @base 🟦
No Bittensor wallet required. Trade Pareton like any other token on Base - through @aeroxyz x @forevermoney_ai, secured by @chainlink CCIP.
Access was the bottleneck, not the tech. The leading inference optimiza
Same model. Same GPUs. Up to 54% more tokens/sec.
We tested the top miner for Qwen3.8-27B against the baseline using @nvidia 's AIPerf: all 600 requests completed, with 23–54% higher output throughput across tested loads.
Every token costs money. Pareton is driving that cost do
We’ve submitted a CoinGecko Token Info Update request for Pareton (SN10).
GeckoTerminal: https://t.co/sMLgoEwRFX
Telegram contact: @xavierlyu
This post confirms that the contract and pool addresses in the request is official.
More GPUs don't always mean more usable GPU memory.
A useful example from LLM inference: a workload hit an OOM. The interesting part wasn't simply how much memory was being used, but what was using it and whether that memory could actually be distributed across GPUs.
Let us
Pareton (SN10): Inference Optimization on Bittensor TAO https://t.co/OEKKQQI3h7
Pareton miners found it, vLLM merged it.
An optimization from our Qwen campaign on #Bittensor SN10 is now upstream in @vllm_project: ~4% more throughput at batch 4–8 for Qwen3.8 with MTP speculative decoding.
Open competition → open-source wins.
PR: https://t.co/Cz7NU1pBqB
1/
🚀 New SGLang campaign coming this week
Launching before Friday, built around a real production workload for a potential paying customer, bringing SN10 one step closer to its first revenue!
Setup:
• Qwen3.8-27B-FP8
• SGLang
• 4× RTX 5090, TP4
• 262K context
Workload: 32 reques
We just put Pareton's engine on an H200 and beat a B200 on Qwen3.8-27B-FP8.
B200, stock vLLM: 106.2 tok/s
H200, Pareton engine: 252.4 tok/s
2.4x on a cheaper GPU. Same prompt, same output length, batch 1.
Third-party verified: https://t.co/jFt0VWXUOA
Miners just made Qwen3.8-27B-FP8 3.5× faster than stock vLLM end-to-end on 1× H200 (median over SWE-agent traces):
• 727ms → 190ms median request
• 58.7 → 221.8 tok/s per request
• inter-token latency p99: 93ms → 38ms
• requests meeting a strict serving SLA: 3% → 100%
• identical
⛏️ Mining starts tomorrow!
Target: Qwen3.8-27B-FP8 on 1x H200
Fee: 0.1 TAO
Patch the engine. Beat the leader. Take the emission!
Learn more: https://t.co/g5j56e2S8H
Inference is not one-size-fits-all. Different models. Different hardware. Different traffic. Different workloads.
Pareton is building a system to search that space, and build the best engine for you.
September 1, 2026
Bittensor Subnet 10
https://t.co/nrpgwq94hL
🚀 Preparing to submit our request to CoinGecko for Pareton!
🔗 GeckoTerminal URL: https://t.co/sMLgoEwjQp
Telegram contact: @xavierlyu
SELLERS DRAINED
85%
28D SELLERS EMPTIED OUT
CAPITULATION
-0.02
BOUNCE COMPOSITE (z)
HEAT 7D
0.17×
VOLUME ÷ POOL DEPTH
HOLDERS IN PROFIT
0%
TRACKED WALLETS · ENTRY BELOW PRICE
WHAT'S CARRYING THE PRICE · 30DMIXEDemission rank #82 · below the bar (rank 32)
30D RETURN
-31.7%
ALPHA vs TAO
WITHOUT CHAIN BUYS
-31.7%
SAME 30 DAYS, PROTOCOL BID REMOVED
CHAIN SHARE OF BUYING
—
CHAIN BUYS vs TRADERS' NET
CHAIN BUYS / DAY
0 τ
PROTOCOL BUYING ITS ALPHA
See who's buying and selling Pareton — every trade classified miner / validator / owner / outsider, plus holder cost basis and the full 23-factor screener.
OPEN THE INSIDER TAPE →