Alibaba announces 2.4T and 27B Qwen3.8 models
Alibaba is preparing to launch new model variants in the Qwen3.8 series next week, featuring a high-capacity 2.4 trillion parameter flagship model alongside a smaller 27 billion parameter variant. The upcoming releases expand the ecosystem of open-weight models, serving high-throughput enterprise infrastructure as well as local developer deployments.
Releasing both a 2.4T mega-scale model and a 27B parameter model simultaneously captures both ends of the AI deployment spectrum—from massive server clusters to local developer workstations.
• The 2.4T parameter architecture pushes the boundaries of open-weight Mixture-of-Experts (MoE) performance.
• The 27B variant delivers a practical balance of high quality and efficient local execution on consumer hardware.
• Offering multiple model scales strengthens competitive pressure on closed proprietary API providers.
DISCOVERED
1h ago
2026-08-03
PUBLISHED
1h ago
2026-08-03
RELEVANCE
AUTHOR
lmstudio