ZCode 3.0
TLDR Dev ● Covered by 2 sources
ZCode 3.0 launched with tiered GLM Coding Plans, from lightweight to high-volume dev work. It bundles deep GLM-5.2 integration and lets you steer coding tasks from chat apps like Telegram.
Z.ai just rolled out ZCode 3.0, and the pitch is less about a flashy new model and more about turning coding assistance into a subscription utility you scale like a phone plan. There are three tiers now: Lite at $16.20 a month for small repos and casual iteration, Pro at $64.80 for daily development on mid-sized codebases, and Max at $144 for teams grinding through large, complex projects. Each tier stacks on top of the last, with Max offering 20 times the usage allowance of Lite and dedicated compute during peak hours.
What's actually new here is the depth of integration with GLM-5.2, Z.ai's flagship model, which now threads through reasoning, code generation, and multi-agent coordination inside ZCode itself. That's not a minor tweak. It means the tool is built to hand off subtasks between agents rather than treating every coding job as one long monolithic prompt. Z.ai is also leaning into something called Goals, a feature for managing long-running tasks that keeps a plan alive through execution and verification instead of dying the moment a session ends.
The more unusual addition is bot control. ZCode can now be started and steered from WeChat, Feishu, or Telegram, which sounds like a small convenience until you consider what it actually implies: developers checking in on a build from a chat thread instead of a terminal window. Given how much of Asia's dev and ops culture already lives inside WeChat and Feishu, this isn't a gimmick so much as a deliberate distribution play.
Z.ai says ZCode plugs into more than 20 coding tools, and pricing undercuts what US-based coding assistants typically charge for comparable usage tiers. Installers are out now for macOS and Windows, with rolling or priority access to newer GLM releases baked into the paid tiers depending on how much you're willing to spend.
My take
I'll say it plainly: the chat-app control feature is the real story here, not the pricing tiers everyone will focus on. Western coding tools keep bolting agentic features onto IDEs while Z.ai is quietly building for a world where WeChat and Telegram are the actual command line. If GLM-5.2 holds up to its coding claims, this is a much bigger threat to Western tooling incumbents than another benchmark chart would suggest.
Read more about this at: TLDR Dev
