Large-context multimodal model
The page describes Kimi K3 as a 2.8T-parameter model with native vision and a 1-million-token context window, built for long-horizon coding, knowledge work, and reasoning.
Kimi K3
Kimi K3 is Moonshot AI’s frontier model for coding, knowledge work, and reasoning. It is available across Kimi.com, Kimi Work, Kimi Code, and the Kimi API, with a 1-million-token context window and native vision support.
Open frontier model for coding and knowledge work
Kimi K3 is Moonshot AI’s frontier model launched through Kimi’s product suite. The launch post describes it as a 2.8T-parameter model with native vision capabilities and a 1-million-token context window, and says it is the first open 3T-class model.
The model is available across Kimi.com, Kimi Work, Kimi Code, and the Kimi API. Across the source pages, Kimi K3 is positioned for long-horizon coding, knowledge work, and reasoning rather than as a single-purpose chat model. Kimi Work uses it for desktop automation and agent workflows, while Kimi Code presents it as a code agent for developer tasks.
Core capabilities
Developer-oriented code agent
Kimi K3 is presented as the model behind Kimi Code, where it is positioned for coding tasks, terminal tool use, and work across large repositories.
Local desktop workflow automation
Kimi Work is described as a desktop local agent that mounts local folders, runs Python in the background, and works continuously on the desktop.
Browser automation with WebBridge
WebBridge lets Kimi Work browse like a human, move across tabs, and extract information from websites for multi-step web tasks.
Scheduled automation
The product page says Kimi Work includes a Cron engine for LLM agent calls, Python/Shell execution, and recurring tasks that can run on a schedule.
Multi-agent research and office output
Kimi Work includes Swarm-based coordination of multiple specialized agents, then turns completed research into PowerPoint decks or Excel sheets.
Practical use cases
Developer coding workflows
Use Kimi Code when you need an AI coding assistant that can work inside terminal and IDE-oriented workflows, help with large repositories, and support programming tasks with a long context window.
Desktop knowledge work
Use Kimi Work for desktop tasks that combine local files, browser actions, and background scripts, such as finding documents, summarizing them, or running repeated office automation.
Automated web research
Use the browser automation layer when a task requires web research, tab navigation, data extraction, or copy-to-Excel style workflows without manual clicking.
Recurring automation
Use Kimi Work’s scheduled tasks when you need recurring LLM calls or Python/Shell jobs to run on a daily, hourly, or conditional basis.
Long-form research and analysis
Use Kimi K3 for long-horizon reasoning tasks that mix code, visuals, and research, such as scientific workflows or interactive reports.
Pros and Cons
Pros
- 1M-token context supports long documents, large codebases, and extended workflows.
- Native vision capabilities let the model work with screenshots and other visual inputs.
- Kimi Work combines local files, browser automation, background execution, and scheduled tasks in one desktop agent.
- Kimi Code is presented as a developer workflow tool with CLI support and a one-subscription model.
- The product suite spans web, desktop, coding, and API access.
Cons
- The launch post says Kimi K3 still trails the most powerful proprietary models overall, even though it performs at a frontier level in the reported evaluations.
- Pricing, plan structure, and enterprise terms are not detailed in the provided pages.
- The coverage provided here is stronger on product positioning than on complete integration and setup documentation.
FAQ
What is the difference between Kimi Work and the web version?
Kimi Work is the desktop agent for deeper workflows: it mounts local folders, can navigate the web through WebBridge, runs Python code in the background, and supports scheduled tasks. The web app is described as better for quick chat and queries.
How does Kimi Work handle local-file privacy and permission?
The page says Kimi Work gives you control over your files and uses an “Ask before acting” safeguard. It prompts for explicit authorization before modifying, overwriting, or running code in local directories.
What can WebBridge do?
WebBridge is Kimi Work’s browser-automation layer. It can browse across tabs, click, scroll, extract data, and complete multi-step web tasks such as checking news, summarizing pages, or scraping stock data into Excel.
What kinds of scheduled tasks does Kimi Work support?
Kimi Work’s Cron scheduler supports LLM agent calls, Python/Shell executions, and other scheduled automation. The page says tasks can run daily, hourly, or conditionally, and that you can use the Keep Computer Awake option if you want overnight execution.
Where is Kimi K3 available?
Kimi K3 is available on Kimi.com, Kimi Work, Kimi Code, and the Kimi API. The launch post says full model weights will be released by July 27, 2026, but it does not describe separate model plans or pricing on the source pages provided.
Quick Facts
- Category
- AI model / productivity / developer tool
- Source domain
- kimi.com
- Primary platforms
- Web, desktop, CLI, API
- Context window
- 1 million tokens
- Notable workflow
- Coding, knowledge work, browser automation, scheduled tasks
- Availability note
- Kimi K3 is listed as available on Kimi.com, Kimi Work, Kimi Code, and the Kimi API
Kimi K3 Alternativen
ByteAsk
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repositories and verifies changes with the real compiler, debugger, sanitizers, and tests before showing a diff. It offers a free tier plus paid plans, with editor connectors and zero-retention handling described in the source.
Ghost
Ghost ist ein terminalbasierter KI-Assistent für Chats, Code-Generierung und Aufgaben direkt in der Kommandozeile. Mit kostenlosen Modellen, Linux, macOS und Windows. Open Source.
CreateOS Sandbox
CreateOS Sandbox is an isolated compute environment for running code and agent workloads inside Firecracker micro-VMs. It is designed for workflows that need machine-level isolation, private networking between sandboxes, and programmatic control through SDK, CLI, or MCP.
hob
hob is an independent workspace for coding agents that keeps agent sessions, terminals, history, and follow-up work organized around the tools and providers you already use. It is aimed at developers who want local control over routing, history, and workspace structure rather than a bundled model stack.
Manta AI
Manta AI is an autonomous web app testing tool for teams that want to map application behavior, catch regressions, and generate tests without writing scripts or maintaining selectors. It works from a URL and supports plain-English test flows, run results with screenshots, and scheduled or deployment-triggered checks.
Redline
Redline is a budgeting tool for Claude Code that paces sessions to stay within time, token, cost, or plan-percentage limits. It uses Claude Code’s native hooks and statusline to help sessions finish with a usable result instead of stopping abruptly.