Text Generation

A natively multimodal general model, with text, image and audio through one door

The model family that made the open-weight route mainstream

An open-weight family spanning many sizes, with multimodal versions

An open-weight MoE with 671B parameters, activating 37B per token

The flagship commercial model from a European open-weights lab

A natively multimodal general model built for very long context

A general chat model known for long context and safety alignment

An open-weight reasoning model with hybrid attention and a million-token context

A reasoning model trained with RL on chains of thought, its weights open under the MIT licence

System-level AI that splits work between on-device and private cloud

A commercial model built for retrieval augmentation and tool use

An open-weight model that mixes a state-space model with a Transformer

Databricks’ open-weights MoE language model

One chat entry point that gathers search, office apps and a multimodal model

A chat model tied to a social platform’s live data

A bilingual Chinese–English open-weight model with very-long-context versions

A Chinese chat assistant known for long-context handling

Tencent’s general model family, with open-weight versions

Conversational AI woven into the operating system and office apps

ByteDance’s general chat model and application

Small, efficient models built from curated data

A general model family aimed at multimodality and on-device use

An open-weight general model aimed at Chinese-language use

A Chinese general model that began with autoregressive blank-infilling pretraining

The chat window that put a large language model in everyone’s hands

Role-play chat with characters you define and talk to over time

A Chinese model that began with knowledge-enhanced pretraining, an early landmark version