Uploaded July 2026 | Updated November 2026, 1 day ago
Z.ai's GLM 5.2 is an open-weight, MIT-licensed model with a one million token context window that's scoring close to Opus and GPT 5.5 on terminal-bench. This video walks through setting it up with dcode, the open source, model-agnostic coding agent, and shows how to trace every step of the agent's behavior through LangSmith.
0:00 Why everyone is talking about GLM 5.2
0:50 The benchmark numbers
1:11 Why a great model still needs the right agent
1:55 Meet dcode, the open source coding agent
2:23 Installing dcode and connecting GLM 5.2
2:57 Kicking off a real coding task
3:25 Authenticating integrations and exploring slash commands
4:00 Checking in on the running agent
4:15 Inside LangSmith: tracing dcode step by step
4:54 Drilling into tool calls and token usage
5:27 Where open models and dcode go next
Extra resources:
Deep Agents: langchain.com/deep-agents
Z.ai's GLM 5.2 is an open-weight, MIT-licensed model with a one million token context window that's scoring close to Opus and GPT 5.5 on terminal-bench. This video walks through setting it up with dcode, the open source, model-agnostic coding agent, and shows how to trace every step of the agent's behavior through LangSmith.
0:00 Why everyone is talking about GLM 5.2
0:50 The benchmark numbers
1:11 Why a great model still needs the right agent
1:55 Meet dcode, the open source coding agent
2:23 Installing dcode and connecting GLM 5.2
2:57 Kicking off a real coding task
3:25 Authenticating integrations and exploring slash commands
4:00 Checking in on the running agent
4:15 Inside LangSmith: tracing dcode step by step
4:54 Drilling into tool calls and token usage
5:27 Where open models and dcode go next
Extra resources:
Deep Agents: langchain.com/deep-agents










