What is Kimi K2.5? Kimi K2.5 is an open-source multimodal large language model released by Moonshot AI in January 2026. With 1 trillion total parameters and 32 billion activated parameters, it outperforms Claude Opus 4.5 in multiple benchmarks, especially in coding and agentic tasks.

OpenCode, the most popular open-source AI coding assistant, now officially supports Kimi K2.5. Even better, you can experience this powerful model for free through various channels. This guide covers configuration methods and usage tips.

The Fastest Triple Rebrand in Open Source History

In the history of open-source software, few projects have undergone such a dramatic rebranding as this AI assistant. In less than a week, it went from Clawdbot to Moltbot, and finally decided on OpenClaw—dubbed by the Reddit community as the “fastest triple rebrand in open source history.”

On January 30, 2026, founder Peter Steinberger officially announced that the project was finally named OpenClaw, ending the naming saga that lasted for several weeks.

TL;DR: Getting “This version of Antigravity is no longer supported” error? This guide provides solutions for official client updates, third-party tool configurations, and Linux/WSL2 environment fixes.

The Problem

When using Google Antigravity (Google’s “agent-first” AI IDE released in November 2025), you might encounter this error:

This version of Antigravity is no longer supported.
Please update to receive the latest features!
Build • gemini-3-flash-preview • interrupted

This indicates that your current client version is outdated, and Google has forcibly disconnected support for older versions on the server side.

Preface: From Burnout to Viral — A Developer’s AI Awakening Journey

Clawdbot (recently renamed Moltbot) has sparked unprecedented buzz in the developer community. The project’s creator, Peter Steinberger, is an Austrian developer and founder of the renowned PDF toolkit PSPDFKit.

After the company received over €100 million in investment from Insight Partners in 2021, Peter completely burned out and disappeared for three years.

In November 2025, he vibe-coded Clawdbot in just 10 days. Weeks later, the GitHub star count soared to nearly 90,000, with a growth curve described as “an unprecedented straight line.” Even more remarkably, Cloudflare’s stock price surged 14% pre-market as developers used it to deploy Clawdbot. People on Instagram who never followed tech started posting photos of themselves buying Mac Minis at Apple stores.

Moltbot (formerly Clawdbot) is a powerful open-source AI assistant that can run locally and interact with you via various messaging platforms (like Telegram, Discord).

This article demonstrates the simplest script-based installation method, teaching you how to integrate it with Telegram while using Google Gemini’s free quota to enhance the experience.

1. Environment Preparation

Before starting the installation, please ensure your server or local environment meets the following minimum requirements. Although the script handles some dependencies, pre-checking avoids many errors.

In the previous article, we discussed how to choose a suitable VPS for Clawdbot (now named Moltbot). Today, we will officially get our hands dirty and teach you step-by-step how to deploy this powerful AI assistant to your server and interact with it anytime, anywhere via Discord.

This article is compiled based on practical sharing from AppSail, covering the complete “one-stop” process from server purchase to security hardening.

What is Moltbot?

Moltbot (formerly Clawdbot) is an open-source AI assistant developed by Peter Steinberger. Unlike commercial products like Claude Code, Moltbot runs on hardware (or VPS) controlled by you, connecting browsers and messaging platforms (such as WhatsApp, Telegram, Discord, Lark) via WebSocket architecture.

As Clawdbot (now renamed Moltbot) continues to gain popularity, many developers and AI enthusiasts are starting to look for suitable deployment environments. Should you invest in high-performance local hardware or choose a flexible cloud VPS? This has become a choice everyone faces.

This article, referencing practical experience from tech blogger AppSail, outlines the best low-cost schemes for deploying AI assistants in 2026.

Local Hardware vs. Cloud VPS: Crunching the Numbers

For many geeks, having a Mac mini M4 as a home server is a dream configuration. However, when we take out the calculator for a detailed check, reality might be a bit “leaner”.

“Google’s being generous again (I advise you stop giving credits, give more Banana quota first)”

For the vast number of AI developers and loyal Google users, there is good news today. Google AI Pro members (including memberships claimed via student verification) can now claim $10 API credit monthly on Google Cloud Platform (GCP).

This $10 might not seem like much, but it’s absolutely sincere for individual developers to test the latest models or run some small Agents. More importantly, this credit can be used to call Google’s most powerful models currently available, including the highly anticipated Banana Pro (Nano Banana Pro) and Gemini 3 Pro.

With Moonshot AI officially releasing Kimi 2.5 and its onboard k0-math reasoning model, Chinese large models have seen a qualitative leap in logical reasoning and mathematical problem-solving capabilities. For developers and geeks, being able to run such a powerful model in a local environment means not only higher data privacy security but also lag-free offline calls.

Although Kimi 2.5 is a massive Mixture-of-Experts (MoE) model, thanks to Ollama’s ecosystem support, we can now easily deploy its quantized version on personal computers.

In today’s fiercely competitive Chinese LLM market, Moonshot AI has once again delivered a blockbuster result.

Recently, Kimi welcomed a comprehensive version iteration (widely known in the industry as Kimi 2.5). This update is not just a stack of parameters, but a qualitative leap in reasoning capability and multimodal understanding by introducing Reinforcement Learning at the underlying architecture level.

1. Core Breakthrough: k0-math Reasoning Model

The most eye-catching focus of this update is undoubtedly the release of the k0-math model. This is Kimi’s first reasoning model adopting the Reinforcement Learning technical route, aiming directly at OpenAI’s o1 series.