chinese-llm

Official Vynex API SDK — one OpenAI-compatible endpoint for GPT/Claude/Gemini. Pay with USDT (no foreign card needed).

Exploring the Frontier of Chinese LLM Writing

This repo introduces MagicData-CLAM, a Chinese SFT dataset, and provides to the community two relevant models that we finetuned. Contact business@magicdatatech.com for more information.

AI Consulting Engine + Smart Routing — Multi-model synthesis decomposes complex questions, routes to optimal models, returns expert analysis. One API key.

A TypeScript agent CLI built around OpenAI-compatible APIs. First-class support for self-hostable and Chinese LLMs (DeepSeek, Qwen, Kimi, GLM, Ollama, MiMo).

OpenAI-compatible code examples for chat, image & video generation using top Chinese AI models (DeepSeek, Qwen, GLM, Doubao, Kimi, MiniMax) through a single API endpoint.

GhostAI 知识库 — 5 分钟拥有专属 AI 知识库 · DeepSeek + bge-m3 · 国内合规部署 · ¥39 起

2026 年大语言模型评测、对比与选型资源整理 / Curated LLM benchmark rankings, comparisons, and selection guides for 2026

🚀 一个 API 连接所有大模型 · Open-source promo site adopted by nexusly.cc · 92+ models · <300ms · up to -50% off · OpenCode/Kiro/Codex 一键兼容

Source-linked index of 69 Chinese AI model releases from 21 labs — parameters, context window, licence and API pricing, each field with its source URL and verification date. CSV/JSON, CC-BY-4.0.

MCP Server wrapping a Supabase pgvector knowledge base (BAAI/bge-m3 embedding) for Claude Desktop / Cursor / Inspector. Tools + Resources + Prompts. Chinese-first.

A reproducible, memory-efficient pipeline for training a 1.2B-parameter bilingual Chinese-English GPT language model from scratch on a single RTX 4090. 一个可复现且显存高效的训练管线，用于在单张 RTX 4090 上从零训练 12 亿参数的中英双语 GPT 语言模型。

The 2026 LLM API price table, rebuilt with one billing convention. 11 Chinese models with input/output/cache prices verified 2026-08-01, why old price tables expire, and how to benchmark your own real cost.

Yisho AI Agent 桌面工作空间：用于编程、办公、产品开发与内容创作，支持国产大模型及兼容 LLM 服务，提供 Windows x64 安装包。

中国大模型API资源大全 - 国产大模型API汇总与对比

语义回响实证数据仓库：多模型对照 + 全流程 7 模式 2026 评测结果 + 论文/图表

国产大模型API Chat Demo - One API for DeepSeek, Qwen, GLM, Kimi. OpenAI-compatible gateway.

🚀 Optimize memory for large language models, enabling 70B models on a 4GB GPU and 405B Llama3.1 on 8GB VRAM without compression techniques.

基于 LLaMA-Factory 的 Qwen2.5-7B 中文指令微调 + Ollama Web 部署完整流程(LoRA + GGUF 量化)

Six Chinese AI models deliver GPT-5.5-level performance at 1/10 to 1/30 the cost. Real pricing data, Artificial Analysis benchmarks, and a practical guide to accessing them from outside China.