Rodney

[Cacheon @cacheon_ai]

We shipped two things over the weekend: a 0.1 TAO miner submission fee and @shadeformai GPU support.

Submission fee: Every on-chain commit now costs 0.1 TAO. Goal is to cut spam and add skin in the game. Fee covers GPU rental first; anything left buys SN14 tokens and burns it. Miner workflow is unchanged. Docs: https://t.co/IfKldbyBTh

Shadeform GPUs: Validator can now pull GPUs from Shadeform alongside @TargonCompute and @lium_io
for evaluations. More supply, less wait time.

Updates on evaluation upgrade coming later this week. Follow our Bittensor Discord channel for ongoing discussions.

[Rodney @TacoShak1]

Whoa - @zcashnames is going to be big. I found the waitlist. Sign up and get alerted before they launch. https://t.co/nqs5zEQl2C

[Cacheon @cacheon_ai]

The first SGLang-based miner hit testnet (5GbtnKr4, UID 15)! This is exactly what Cacheon is designed for: bring any language or framework you want, as long as you match the OpenAI-style API we document (/health, streaming chat completions, logprobs, etc.).

[Tseu Tseu - τao @tseutseutao]

$TAO #SN14 - Subnet teams execution is increasing by the day. Go get them @cacheon_ai!

[wejh @Wejh69]

Agents mining the "agent" subnet

🔥🔥🔥🔥🔥🔥🔥
https://t.co/6Q3oYtziFa
https://t.co/QJLLtbR35D

[Punisher ττ @CryptoZPunisher]

Bittensor >> $TAO >> $dTAO

Subnet 66: Ninja

When the boss is cooking, you know what comes out of the kitchen will earn many stars.
A leader, a founder, a subnet owner who doesn’t sell dreamshe builds seriously.
What he has proven over the years is enough for me to build a position.

[Punisher ττ @CryptoZPunisher]

Bittensor >> relaxation

SN66: Root 66 🥷🙃

I don’t know who to credit for this image, but it’s awesome, thanks for this great moment.
Route 66 by @const_reborn .

[Punisher ττ @CryptoZPunisher]

Bittensor >> $TAO >> $dTAO

Subnet 24: Quasar just announced a partnership with @adaption_ai , and at the same time, the team is stating they aim to deliver state-of-the-art.

That’s a strong ambition.On paper, the move is coherent:

→ evolving architecture
→ compute with Targon
→ now strengthening the data layerThe stack is starting to take shape.

But as always:an announcement is a signal, not proof.The real test will be elsewhere:

→ real performance
→ actual usage
→ adoptionThe ambition is there.

The direction seems clear.

Now it’s time to execute.

[Punisher ττ @CryptoZPunisher]

#Biττensor >> ∆ τ << #τₐcc
> $TAO <

Subnet 24: Quasar
@QuasarModels

Will Quasar be as brilliant as what the team is about to announce?

They’d better be confident, because the community’s reaction will be immediate.

@TroyQuasar
@Farahatyoussef0

[Ridges AI | SN62 @ridges_ai]

⛰️ Ridges Update

Harbor integration is live.

For context, Harbor (by Terminal Bench) is a framework for evaluating agents across more languages and more complex tasks. It keeps our evals aligned with what our agents are actually capable of.

Next up: Synthetic Bench, our generated benchmark designed to test agents on problems outside public datasets. Harder to game, closer to the real world.

[τroy @TroyQuasar]

Thats not it for April

[Quasar @QuasarModels]

🚀 Quasar Competition Starts Now

Miners, this is your moment.

The Quasar codebase is now fully open.
We are launching a global competition to distill knowledge into Quasar and push it to state-of-the-art performance.

[🏆 Winner takes it all ]

The first training run begins now
Anyone can participate bring your best methods, your best ideas.

For details, updates, and how to get started, join the Quasar Discord channel on Bittensor.

Let’s build 👇

[Quasar @QuasarModels]

First announcement of April.

We introduce Quasar-3B (1B Active) a looped continuous-time transformer built for long-context intelligence

Quasar uses a hybrid architecture combining Quasar layers + GLA, enabling efficient stateful reasoning with stable long-range dependency handling.

This release is the base model of the Quasar system, designed for distributed training and distillation at scale. In the coming days, SN24 miners will begin working with us to distill knowledge from frontier models like Qwen into Quasar, pushing toward a new SOTA in long-context modeling.

The training roadmap is staged:

Stage 1: Quasar-RoPE stable pretraining with 16K context, establishing the base representation for the system
Stage 2: Quasar-DroPE continued training with distillation as the core mechanism, removing positional encodings and scaling toward 5M token context

In both stages, distillation remains the central training signal, driven collaboratively through SN24 miners using knowledge transfer from frontier models.

Read more about Quasar and get ready for mining 👇

[Project Nobi @projectnobi_tao]

🔮 Project Nobi Arena: SN62 vs SN66

Disclosure first: Project Nobi is mining SN66 and is going to mine SN62. We run Ridges and Ninja agents. We've studied these subnets' architectures from the inside.

@ridges_ai
@arbos_born

That firsthand experience is why we built the Arena.

#SN62 has some of the most sophisticated coding agents in the ecosystem — structured problem decomposition, hypothesis tracking, multi-step verification before finishing. But right now, they're only benchmarked internally. There's no external, cross-subnet measurement.

The Arena changes that. SN62 agents go head-to-head against SN66 (Ninja) on identical SWE-bench tasks : same repos, same constraints, deterministic pass/fail scoring. No subjective evaluation. Patches either pass the tests or they don't.

We've watched SN66 agents improve for months. We know what they're capable of. But capability without external benchmarks is just a story. The Arena puts Ninja agents head-to-head against SN62 (Ridges) on identical SWE-bench tasks : same repos, same constraints, same deterministic scoring. Every patch is public. Every result is verifiable.

The format: 50 matches. 5 agents per side. Round-robin, home and away. 1 match per day, 7 weeks.

Our commitment: We mine both subnets. Scoring pipeline is open. Our miners compete under standard rules. Results published raw before any commentary.

We have skin in both games. That's why we can referee this one honestly.

Season 1 in ~3 weeks.
Full rulebook: https://t.co/T7VPyVVvdm
Scripts & verification: https://t.co/VABgq2gbv2 🏆

#Bittensor @opentensor $TAO
@const_reborn @JosephJacks_

[Punisher ττ @CryptoZPunisher]

#Biττensor >> ∆ τ << #τₐcc
> $TAO <

Subnet 66: ninja

➡️https://t.co/Lqye4g6go8

An outer agent (Arbos) orchestrates this loop continuously. It:

▫️Analyzes rollouts to find failure patterns
▫️Writes Rust patches to Ninja's source code
▫️Rebuilds the binary, re-runs tasks, measures the delta
▫️Pushes improvements that show net-positive results, reverts the rest

➡️https://t.co/l3smpwknrS

➡️https://t.co/hYwPYHQ291
➡️@arbos_born

I just opened the SN66 dashboard.
This isn’t a subnet.
It’s an arena.
A king.
Challengers in queue.
Hundreds of duels.
Thousands of rounds.
Agents are no longer benchmarked.
They are battle-tested.

[Ridges AI | SN62 @ridges_ai]

Ridges isn’t going anywhere.

Ridges is bittensor, bittensor is Ridges.

We’re aligned. We're here to build and to keep raising the bar.

More updates soon.

[Jesus Martinez @JesusMartinez]

Warning: Do NOT Buy SN3 “Templar” right now

@const_reborn warns any buys now can give Sam more funds to dump on you

Wait until they sort it out &amp; gain control again

[const @const_reborn]

https://t.co/IdB9DPr5u0

[TAO Flows @TAOFlows]

🚨Hey @AltcoinDaily!

How many likes for you to do the @const_reborn interview?

Seems like good timing don’t you think?
$TAO

[Punisher ττ @CryptoZPunisher]

$TAO

He is “dead”… yet still the leader of his category.

Now close your eyes and think about this:

In the age of instant information, most information, at every level, is just noise.

The best example? Donald Trump

He has mastered the art of capturing attention and manipulating the masses.

Exactly like certain influencers chasing views:
amplify the noise, create the show, divert attention.

The spectacle attracts weak minds.
Chaos sells.
Meanwhile, smart money does not react to noise.

It profits from your anxiety… amplified by social media.