Academic and Industry Research

Doubao's continuously updated model endpoint for long-running agents, coding, and multimodal understanding.

Updated 2026-10-09

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

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

Updated 2026-10-09

Google's multimodal model for complex reasoning, coding, and large-scale document analysis, with roughly 1M-token input support.

Updated 2026-08-29

OpenAI's deep-reasoning model for math, science, and complex planning.

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

An open model for code generation, completion, and repository-level understanding.

Updated 2026-08-29

A Chinese deep-learning textbook from Li Mu's team balancing code and theory.

Updated 2026-07-09

系统整理提示词技术、上下文学习、推理方法与安全实践。

Updated 2026-09-03

李沐的中文论文精读与深度学习课程。

Updated 2026-09-03

The MoE model that drew attention with exceptionally low inference cost.

Updated 2026-08-29

A representative open vision-language model from the Chinese ecosystem.

Updated 2026-08-29

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

Updated 2026-08-29

总结Prompt&LLM论文，开源数据&模型，AIGC应用

Updated 2026-09-03

📚 Two books on harness engineering — the design philosophies behind Claude Code & Codex: constraints, query loops, context governance, mult

Updated 2026-09-03

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

Updated 2026-08-29

ByteDance's Doubao model with a broad multimodal and video-generation ecosystem.

Updated 2026-08-29

An open ASR model for Chinese, multilingual, and long audio.

Updated 2026-08-29