inclusionAI Releases Ling 3.0 flash Hybrid-Linear Reasoning Model

AdinaYakup · x · 2026-08-05

inclusionAI has released the new Ling 3.0 flash model, designed to deliver strong reasoning performance with minimal compute.

The model has 124B total parameters with only 5.1B active parameters (MIT license). The team claims its performance matches flagship models with 1T active parameters. Additionally, thanks to smart caching, it achieves 60-80% faster response times on long inputs compared to traditional architectures.

Related event: Ant Group open-sources Ling-3.0-flash: 124B params, only 5.1B active(11 posts)→

Original post →

More from Models

Models channel →