Multimodal model achieving SOTA performance
From Images to High-Fidelity 3D Assets
Qwen3-omni is a natively end-to-end, omni-modal LLM
MapAnything: Universal Feed-Forward Metric 3D Reconstruction
State of the art LLM and coding model
MiniMax-M2, a model built for Max coding & agentic workflows
Large-scale autoregressive pixel model for image generation by OpenAI
Code for the paper "Improved Techniques for Training GANs"
High-performance MoE model with MLA, MTP, and multilingual reasoning
VaultGemma: 1B DP-trained Gemma variant for private NLP tasks
Qwen3-Next: 80B instruct LLM with ultra-long context up to 1M tokens