kimi-k3

Swap GPT for any LLM by changing a single line of code. Xinference lets you run open-source, speech, and multimodal models on cloud, on-prem, or your laptop — all through one unified, production-ready inference API.

A 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU.

A powerful Zotero AI and MCP plugin with ChatGPT, Gemini 3.7, Claude Fable 5, Claude Opus 5, DeepSeek V4, Grok, OpenRouter, Kimi k3, GLM 5.3, SiliconFlow, GPT-oss, Gemma 4, Qwen 3.8

Battle-tested Claude Code, OpenAI Codex, Cursor configs, plugins, hooks and agents with Kimi, MiniMax and GLM API support.

ServerKit is a lightweight, modern server control panel for managing web applications, databases, and services on your VPS or dedicated server. Built with Python/Flask backend and React frontend.

🚀 Pytorch Distributed native training library for LLMs/VLMs with OOTB Hugging Face support

Run full Kimi K3 on a single device. And an OpenAI-compatible API server for local chat and coding agents.

Long Horizon Terminal Benchmark with Dense Reward Grading

eLLM: Run Long-Horizon Inference Faster on CPUs Than on GPUs

11 installable skills for Claude Code, OpenAI Codex, and Hermes, for finding your unknowns before they get expensive: blindspot pass, brainstorms, interviews, references, plans, notes, pitches, a merge quiz, plus a context audit, agent interface design, and progressive disclosure. Community distillation, not official Anthropic.

A WIP vibe friendly game engine.

11 installable editorial systems for AI agents, based on leading public style guides.

Kimi K3 Code Free desktop - kimi k3, is kimi k3 free, kimi k3 huggingface, kimi k3 open weights, kimi k3 benchmarks, kimi k3 vs fable 5, kimi k3 cost, kimi k3 reddit, chinese ai, 200K context, GitHub PR review, Moonshot. Install:🡇

Autonomous multi-agent pentest framework — plans, exploits, verifies (proof-required), CVSS-scores and writes client-ready reports. 104/104 (100%) on the XBOW validation benchmarks, powered by Kimi K3.

Kimi K3, deepseek v4 flash, glm 5.2, qwen 3.8 27b, Muse Glimmer, Nemotron, glm 5.3

SillyTavern extension: client-side reasoning_content thinking prefill for Kimi/Moonshot models (kimi-k3-jb patch equivalent, no server modifications)

AI驱动的影视广告创作平台/AI-driven platform for film and television advertising creation

Modern desktop client for FTP, SFTP, SSH, and S3-compatible storage. FTP, SFTP, SSH, and S3 in one clean workspace.

CPU inference for Kimi K3, a 2.78T-parameter MoE LLM, in pure Rust. No GPU, no BLAS, no PyTorch. Streams the checkpoint from disk. Byte-identical port of kimi-k3-in-c.