Quasar (@QuasarModels) / X

Quasar

312 posts

Quasar

@QuasarModels

Bittensor subnet built to crush the long-context barrier | SN24 | Owners : @farahatyoussef0 and @troyquasar Backed by @const_reborn

SN24

Joined November 2025

Pinned
Quasar
@QuasarModels
20h
Quasar is entering its next chapter on Bittensor SN24. We are moving toward a 10T-token decentralized training run. The idea is simple Quasar Models needs more useful training, not just bigger parameter counts. Real model quality comes from tokens, data quality, training
23K
Quasar reposted
TAO Protocol
@taoprotocolonly
11h
Most large language models are surprisingly forgetful. @QuasarModels (SN24) — built by SILX AI — is trying to fix that at the architecture level: linear-time attention, ~2M-token context, no fragile position embeddings. Memory at the protocol layer. → taoprotocol.org/quasar-sn24-th…
1.1K
Quasar reposted
TAO.com
@TAO_dot_com
19h
TEN TRILLY: Quasar moving toward a 10T-token decentralised training run on Bittensor SN24. Phase 1: 5T tokens to produce a stronger checkpoint. Phase 2: another 5T tokens on top. If successful, this would be one of the largest token-scale training runs in decentralised AI.
Quasar
@QuasarModels
20h
Quasar is entering its next chapter on Bittensor SN24. We are moving toward a 10T-token decentralized training run. The idea is simple Quasar Models needs more useful training, not just bigger parameter counts. Real model quality comes from tokens, data quality, training
2.8K
Quasar reposted
hendrik
@Farahatyoussef0
19h
A 10T-token run has never happened in the history of decentralized AI But Quasar is special. Quasar has its own unique architecture, and we are nerdy enough to achieve such a task We’ll share how we made this possible and exactly when it begins. Keep an eye out
Quasar
@QuasarModels
20h
Quasar is entering its next chapter on Bittensor SN24. We are moving toward a 10T-token decentralized training run. The idea is simple Quasar Models needs more useful training, not just bigger parameter counts. Real model quality comes from tokens, data quality, training
1.6K
Quasar
@QuasarModels
20h
Replying to @QuasarModels
training.silxinc.com Keep an eye on our account and Discord We will be sharing updates as the training system rolls out It will not become effective immediately. First, we will share the design, roll out the system, run tests, and then fully switch over once everything is
918
Quasar reposted
hendrik
@Farahatyoussef0
Jun 8
Let’s fucking go
Quasar
@QuasarModels
Jun 8
Today we’re releasing Quasar-Preview! Our first public proof that the Quasar architecture works at real scale. [ 18B MoE - 2B active / 5M context ] Built with Loop Transformer + Quasar attention Trained on Bittensor through decentralized infrastructure 👇
2.7K
Quasar reposted
τroy
@TroyQuasar
Jun 8
Going to sleep now. I’ll be uploading a lot of updates and repos for the model and subnet. Have a good night, everyone.
2K
Quasar reposted
τroy
@TroyQuasar
Jun 8
This is a very early Quasar model, trained on only 0.1% of our full token budget, so there is still much, much more room to improve!
Quasar
@QuasarModels
Jun 8
Today we’re releasing Quasar-Preview! Our first public proof that the Quasar architecture works at real scale. [ 18B MoE - 2B active / 5M context ] Built with Loop Transformer + Quasar attention Trained on Bittensor through decentralized infrastructure 👇
4.8K
Quasar reposted
const
@const_reborn
Jun 8
5M context length agentic model from Quasar.
Quasar
@QuasarModels
Jun 8
Today we’re releasing Quasar-Preview! Our first public proof that the Quasar architecture works at real scale. [ 18B MoE - 2B active / 5M context ] Built with Loop Transformer + Quasar attention Trained on Bittensor through decentralized infrastructure 👇
17K
Quasar
@QuasarModels
Jun 8
Replying to @QuasarModels
While the model has seen fewer than 1.5T tokens, it already shows amazing performance and reveals the strength of the new Quasar long-context architecture, supporting up to 5M context. The model is still early let’s push it forward together!
4.9K
Quasar
@QuasarModels
Jun 8
Replying to @QuasarModels
silx-ai/Quasar-Preview · Hugging Face
From huggingface.co
5K
Quasar
@QuasarModels
Jun 8
Today we’re releasing Quasar-Preview! Our first public proof that the Quasar architecture works at real scale. [ 18B MoE - 2B active / 5M context ] Built with Loop Transformer + Quasar attention Trained on Bittensor through decentralized infrastructure 👇
116K
Quasar reposted
Andy ττ
@bittingthembits
Jun 6
🚨 Pay attention to Quasar SN24 $TAO going after one of AI’s biggest structural moats, Compute Efficiency. Most LLMs hit the same wall attention scales horribly with long context, making it expensive and controlled by a few of the biggest clusters. Quasar is destroying that
Quasar
@QuasarModels
Jun 4
SN3 is now training Quasar models. This time, we’re starting with a 10B model built on the new Quasar architecture an architecture we believe represents the future of LLMs, especially for long-context reasoning and memory. Quasar is designed to push models beyond short-window
5.5K