Another model-network kicking off: Ingot (97)
Their latest release Ingot-I quantized Qwen 27B into FP8 better than Alibaba's own quantization in 1 week.
Next target is 2-bit and then the larger GLM models.
Their latest release Ingot-I quantized Qwen 27B into FP8 better than Alibaba's own quantization in 1 week.
Next target is 2-bit and then the larger GLM models.
5
21
182
11,383
Promising mining result from @reaxlabs (92).
They built this mining network yesterday and are beating Perplexity today.
Congrats goes out to the ML miner talent on Bittensor, and the team.
They built this mining network yesterday and are beating Perplexity today.
Congrats goes out to the ML miner talent on Bittensor, and the team.
Meet REAX 1, our open decision model built on Perplexity Decider v1.2.
Our own measurement: +2.4 points over its base, ahead in all 6 test draws and all 10 areas.
Open competition starts today 13:00 UTC:
=> Beat it, earn the rewards
see reax.co/competitions
11
36
190
7,754
const retweeted
New Quantization Service
TPN Bench now offers a full Model Quantization Toolkit, which includes methods like I-matrix and Quantization-Aware Distillation.
Here's how it works:
Users can upload any Hugging Face model and select the quantizations they want;
TPN Bench will tions, upload them to a bucket, and then send back the links to those quants. It can also upload them to Hugging Face if given access to it.
Choose between regular quantization, I-matrix, or more advanced techniques like quantization-aware distillation, which require heavy infrastructure that we run for you.
Simplify the whole quantization model pipeline with TPN Bench!
🔗 bench.trueperformancenetwork.com
TPN Bench now offers a full Model Quantization Toolkit, which includes methods like I-matrix and Quantization-Aware Distillation.
Here's how it works:
Users can upload any Hugging Face model and select the quantizations they want;
TPN Bench will tions, upload them to a bucket, and then send back the links to those quants. It can also upload them to Hugging Face if given access to it.
Choose between regular quantization, I-matrix, or more advanced techniques like quantization-aware distillation, which require heavy infrastructure that we run for you.
Simplify the whole quantization model pipeline with TPN Bench!
🔗 bench.trueperformancenetwork.com
0
4
22
1,682
const retweeted
We just killed cold outreach.
Introducing Leadpoet: the AI sales agent that finds buyers who need you now, writes personalized outreach, and follows up on LinkedIn and email.
Used by Dropbox, OnePatch (YC26), Muttdata.
Comment “Poet” for 100 free lead credits to find qualified buyers.
Introducing Leadpoet: the AI sales agent that finds buyers who need you now, writes personalized outreach, and follows up on LinkedIn and email.
Used by Dropbox, OnePatch (YC26), Muttdata.
Comment “Poet” for 100 free lead credits to find qualified buyers.
510
92
565
121,255
Our servers immediately got destroyed. Thanks everyone. The masses yearn for the Bittensor videos.
3
5
102
3,438
Pretty cool to see a subnet mine the root network for recursive self improvement.
The 88 Edge and the State of Root Reborn
TL;DR: 88 Root is clearly outperforming.
Root Reborn's active weight setting began on Aug 27, followed by active rebalancing. We've compiled a performance chart of 88 Root alongside major root validators, using lifetime performance data tops the chart.
Two starting dates are provided for context. 88 Root began as an sn88 validator on Aug 16 and did not become a root validator across all 128 subnets until Sep 3. Its performance data during the interim period is largely noise from a single subnet.
While we expect 88 Root to continue outperforming, the tight guardrails on active rebalancing leave limited room for short-term performance differentiation. The 2% price band restriction makes it nearly impossible to enter or exit fast-moving subnets. The 10% daily turnover limit means a full rebalance can take days or even weeks. On the other hand, these restrictions do help prevent manipulation and mitigate risk. Root has always been about risk-free yield after all.
As the 88 alpha edge becomes more evident over time, we also expect and welcome healthy competition from fellow validators. Many are clearly run by talented teams.
Meanwhile, the unrestricted alpha edge remains in the 88 Quant Fund. Despite the still-anemic market, we're preparing to reopen the fund soon.
Stay tuned. #SN88 #Bittensor $TAO 🚀
4
15
142
7,145
the grand masters of decentralized training.
In decentralized training, every merge stops the whole swarm from training while it waits on the slowest member.
With our latest update, merging is ~5x faster and training is almost 20% cheaper.
The biggest unlock was to convert the network graph into a tree during merging.
🧵 1/(n)
9
24
212
9,034
const retweeted
Live: Timestral now on Ephemeris.
Released yesterday, available today: @MistralAI’s 209M SOTA time series foundation model is live on our inference platform.
Connect your agent via our MCP, supporting Claude Code, Codex, Hermes and more - get SOTA-level forecasting:
Z
Released yesterday, available today: @MistralAI’s 209M SOTA time series foundation model is live on our inference platform.
Connect your agent via our MCP, supporting Claude Code, Codex, Hermes and more - get SOTA-level forecasting:
Z
2
8
29
3,273
const retweeted
Ingot (SN97) has been live for 4 days.
Our current king is already a better quantization of Qwen3.8-27B than Qwen's own FP8 release: 34% closer to the original model.
Same size, more of the original intact. And this is FP8 only.
Lower bits are next. 👀
theingot.ai
Our current king is already a better quantization of Qwen3.8-27B than Qwen's own FP8 release: 34% closer to the original model.
Same size, more of the original intact. And this is FP8 only.
Lower bits are next. 👀
theingot.ai
3
4
48
2,816
const retweeted
Thanks everyone, what a week!!
I think we may have just proven decentralized training competitions for use case specific decision models is a thing that works...
I think we may have just proven decentralized training competitions for use case specific decision models is a thing that works...
Quick moment to step back, because this was a big week
• 3 days ago REAX went live. Today 217 miner slots serve decision models and 65 independent miner wallets are earning... almost all traffic routed to the network is now answered by your GPUs, with zero failed verification del-training competition has its first two champions
=> demand forecasting (UID 71, 91.1 vs the 38.2 baseline) and
=> risk assessment (UID 251, 64.2 vs 31.7).
Both more than double what we started with.
Both passed a confirmation draw, a head-to-head playoff, and copy/twin checks against every earlier entry.
Rewards start with the next weight sets.
• Many fixes this week came straight from the mining community:
real-speed probes per hotkey, keep-alive for qualified entries, re-registration without copy flags, a fairer small-GPU tier, and Senpaii's confirmation-margin PR.
Next: protection for queued entries, several machines per hotkey, and less registration churn.
Thank you all, especially dear miners, for the hard work, the bug reports and the patience.
A young subnet has rough edges, and you're the ones helping us smooth them out.
Great things lie ahead! 🫡
4
7
51
4,330
Pretty awesome actually. From idea to economy in a few hours. Send your agents to mine this with us.
Now live on #bittensor SN1, incentivized.
Prompted by @const_reborn, built by agents.
Prompt → production in 6 hours, zero human dev time. Next one: <1h.
We’re industrializing incentive mechanism design.
7
24
174
11,250
const retweeted
yet we still are one of the only platforms in the world with access to b300s on-demand; not gated by any sales call, kyc form, or waitlist.
and at the same prices that the big clouds currently charge for 3 month+ long commitments, we only charge you for the time you use it.
and at the same prices that the big clouds currently charge for 3 month+ long commitments, we only charge you for the time you use it.
help.
our service is too good and agents keep renting everything.
we need more supply
GPU owners who want 80% higher payout than reserved contracts, please reach out or add your nodes in 5 min: provider.lium.io
4
44
251
11,233
Null consensus scales Bittensor subnets 10x and allows POW registration (no Tao needed for an agent to mine).
First step adding more freedom and scale to our incentive systems.
First step adding more freedom and scale to our incentive systems.
Bittensor recently released Null Consensus, a major protocol update expanding subnet capacity from 256 to 2,500 participants.
It introduces a new way to determine miner rewards, giving subnet owners an alternative to Yuma Consensus.
26
69
439
32,344
const retweeted
Yesterday engy.ai processed 50B+ tokens in a single day, and we’re just getting started.
check our stats here: engy.ai/usage
check our stats here: engy.ai/usage
13
52
260
16,068
const retweeted
Lium is hands down the best place to put your GPUs.
You get paid for idle compute, 95% of rental revenue, and additional bonuses.
The average time from a gpu added to our network to rented is <1 hour.
and at any given moment 9/10 GPUs on our platform are rented.
You get paid for idle compute, 95% of rental revenue, and additional bonuses.
The average time from a gpu added to our network to rented is <1 hour.
and at any given moment 9/10 GPUs on our platform are rented.
4
47
281
11,422
const retweeted
Introducing DecisionModels @decisionmodels_
I've spent the last four weeks benchmarking decision models (I run JevBench).
Today I'm launching the place to actually use the best ones:
decisionmodels.io
➡️Top open models
➡️Up to 4x faster on our own REAX engine
➡️From ens.
Expect more reveals soon, and read 👇
I've spent the last four weeks benchmarking decision models (I run JevBench).
Today I'm launching the place to actually use the best ones:
decisionmodels.io
➡️Top open models
➡️Up to 4x faster on our own REAX engine
➡️From ens.
Expect more reveals soon, and read 👇
71
78
492
1,098,082