Quasar subnet 24
Quasar focuses on long-context AI. Its official repository describes a Bittensor competition in which miners train Quasar-compatible language models, publish model revisions, and are evaluated by validators. The subnet's core editorial angle is overcoming context-window limitations through open competition around long-context model capability.
About Quasar subnet 24
Quasar is SILX Labs' long-context language-model subnet for competitive model training and evaluation.
Language models struggle with very long contexts, high memory costs, and reliable evaluation across extended inputs.
LLM researchers, long-context application builders, model miners, validators, and AI infrastructure analysts.
Team and ownership
Quasar is developed by SILX AI. Eyad Gomaa is a SILX AI co-founder and leader associated with Subnet 24, and Ahmed Taha appears as a SILX team member in public profiles.
- Eyad Gomaa, Co-Founder / CEO, SILX AI; Quasar lead: AI researcher and entrepreneur working on advanced AI architectures.
Public partnerships
- Bitstarter: Bitstarter supports Quasar through ecosystem launch support rather than a deeply documented commercial partnership.
Key pages:
learn about Bittensor,
Bittensor glossary,
Bittensor FAQ,
Bittensor subnet screener,
portfolio tracker,
subnet baskets,
Neuralteq validator,
and fees.