Local Deployment

Kimi's new multimodal open-weight flagship for long documents and coding agents.

Updated 2026-10-09

Qwen's new unified image model for generation, editing, and native transparency under a research license.

Updated 2026-10-09

An open 27B vision-language model for self-hosting, with optional context extension.

Updated 2026-10-09

Qwen's large open text weights, with 2.4T total parameters and 262K native context.

Updated 2026-10-09

An open-weight preview of the Qwen4 architecture with vision and 262K native context.

Updated 2026-10-09

DeepSeek's open general model for cost-effective chat and coding.

Updated 2026-08-29

Alibaba's open model family with many sizes and strong Chinese and multilingual ability.

Updated 2026-08-29

A broad open family balancing Chinese, multilingual, and coding ability.

Updated 2026-08-29

Moonshot's open trillion-parameter MoE model built for agent capabilities.

Updated 2026-08-29

Qwen's closed flagship for high-end hosted experiences.

Updated 2026-08-29

Qwen's open speech synthesis and voice-cloning model.

Updated 2026-08-29

Meta's open multimodal MoE model family for private deployment.

Updated 2026-08-29

One of the first 6B open models widely adopted by the Chinese community.

Updated 2026-08-29

微软云上部署 OpenAI 模型的官方文档。

Updated 2026-09-03

01.AI's open-weight Chinese and English model.

Updated 2026-08-29