Tag
An early hands-on look at Ling 3.1 Flash, a 560B-parameter MoE model with 25B active parameters coming soon, where the author built an intricate musical instrument using it and teases what others will create on release.
Tencent has previewed full benchmarks for its Hy4 model, providing detailed performance assessments and insights into its capabilities.
mindlab-research releases Macaron-V1-Preview-749B, a 749 billion parameter large language model, available on Hugging Face.
Despite Anthropic stating that its powerful Mythos model would remain restricted, users have repeatedly spotted 'Mythos 1' in the Claude Code interface and in source code, suggesting an unannounced staged rollout or quiet production testing.