Alternatives to Kimi K

Open-weight agentic coding models — Laguna XS 2.1 (33B) and Laguna S 2.1 (118B) — built for code that cannot leave your security boundary.

Imbue is an open AI lab publishing modular, open-source coding-agent tools you run and inspect yourself.

Falcon LLM delivers Apache-2.0 open-weight language and multimodal models you can self-host, from hybrid Transformer-Mamba to Arabic ASR

Hermes Desktop is an open-source desktop GUI for the Hermes Agent, with a built-in learning loop that turns your tasks into reusable skills

Open-source library of 817 structured cybersecurity skill files your AI coding agent loads on demand, mapped to MITRE ATT&CK and licensed Apache 2.0.

Frontier code models built for ultra-long-context software engineering and AI research automation.

Open-weight agentic coding models you can self-host, from 33B on-device to 118B frontier-class.

Open-source terminal coding agent that runs against any model you point it at — OpenAI, Gemini, Codex, Grok, Ollama, LM Studio and 200+ more.

Open-source desktop workspace that runs Claude Code, Codex, OpenClaw, and other CLI coding agents side by side with per-task model routing.

Open-source SuperAgent harness that researches, codes, and creates inside a persistent Docker sandbox — MIT licensed and self-hosted.

Open-weight 35B Mixture-of-Experts model with ~3B active parameters for local agentic coding and reasoning.

AppWorld is a simulated-world benchmark for evaluating AI coding agents across 9 apps and 457 APIs.

Open-source CLI agent harness that turns one prompt into a graph of specialized agents across Claude, GPT, Kimi, Grok, and GLM.

Xiaomi's free, open-source terminal AI coding agent with unlimited context and the 309B-parameter open-weight MiMo-V2-Flash model.

Open-source frontier models for domain knowledge work, trained with on-policy reinforcement learning.

Cosine Genie is a coding agent built on the Lumen model family, trained on real production code and deployable air-gapped inside your own perimeter.

OpenAI's cloud coding agent that runs many software engineering tasks in parallel, each in an isolated sandbox preloaded with your repo.

Open-source agentic coding assistant that runs inside your terminal and wires any LLM into your CLI workflow.

Open-source coding agent that reads your repo, edits files, and runs terminal commands across VS Code, JetBrains, CLI, and a desktop app.

Open-source multimodal model with native function calling, built for agents that both see images and take action on them.

Kilo is an open-source AI coding agent with 500+ models across VS Code, JetBrains, CLI, cloud, and Slack — at exact provider rates.

MIT-licensed terminal coding agent that reads files, edits code, and runs commands with the model you choose.

Open-source AI coding agent that runs in your terminal, IDE, or desktop app and connects to 75+ LLM providers.

Liquid AI's open-weight LFM2.5 model family runs native text, vision, and audio AI locally on CPU, GPU, or NPU.

DeepSeek-only coding agent you run locally, with plan approval, per-turn checkpoints, and a cache-first loop built for hours-long sessions.

Collaborative AI agent that builds apps, decks, docs, and designs on its own cloud computer while you watch and steer.

Herm is an open-source, containerized coding agent that runs every terminal session in its own sandbox, so the agent acts without touching

Devstral 2 is Mistral's open-weight coding model family — a 123B and 24B pair, plus the Apache 2.0 Mistral Vibe CLI for autonomous software

Open-source 27B agentic coding model with multimodal reasoning and a 50%-leaner ThinkingCap fine-tune.

Zhipu GLM (Z.ai) ships the open-weights GLM-5.3 family, full-modality MaaS APIs, and autonomous agents like AutoGLM and GLM-PC.