Thinking Machines Lab releases Inkling 975B MoE model
Inkling is a 975B parameter sparse Mixture-of-Experts model featuring short convolutions, embedding RMSNorm, and relative-position bias. The open-weight release includes a detailed benchmark profile and architectural analysis.
1 source
Daily brief
Get tomorrow's AI brief in your inbox
More stories today
- Onton releases Ontology 1, a neurosymbolic search model
- Agentic SOC Platform uses AI agents for security triage
- WiFi-3D-Fusion performs real-time 3D human pose estimation
- Eight AI agents automate Obsidian vault
- Comfyanon says H3 will still release