Z.ai coding model connection issue

[chetverovod]

Question

OS Ubuntu 22.04.5 LTS
opencode ver. 1.0.65

Hello! I have an issue in Z.ai model using in opencode. I have an Z.ai API-key. It is added to opencode by *auth login * command. Access to model works okey when I make request by curl:

curl -X POST "https://api.z.ai/api/coding/paas/v4/chat/completions" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer  my api-key" \
-d '{
  "model": "glm-4.6",
  "messages": [{"role": "user", "content": "Write a short story about a robot"}],
  "stream": true
}'

request and answer log:

 curl -X POST "https://api.z.ai/api/coding/paas/v4/chat/completions" \ST "https://api.z.ai/api/coding/paas/v4/chat/completions" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer 6ebc214feae84a9899e35a5726930070.eeVimJD98DJYlxm2" \
-d '{
  "model": "glm-4.6",
  "messages": [{"role": "user", "content": "say Hello!"}],
  "stream": true
}'
data: {"id":"2025111517384495eaf2ee418443b1","created":1763199524,"model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"","reasoning_content":"\n"}}]}

data: {"id":"2025111517384495eaf2ee418443b1","created":1763199524,"model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"1"}}]}

data: {"id":"2025111517384495eaf2ee418443b1","created":1763199524,"model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"."}}]}

data: {"id":"2025111517384495eaf2ee418443b1","created":1763199524,"model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","content":"","reasoning_content":" "}}]}

data: {"id":"2025111517384495eaf2ee418443b1","created":1763199524,"model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" **"}}]}

data: {"id":"2025111517384495eaf2ee418443b1","created":1763199524,"model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"An"}}]}

data: {"id":"2025111517384495eaf2ee418443b1","created":1763199524,"model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"alyze"}}]}

data: {"id":"2025111517384495eaf2ee418443b1","created":1763199524,"model":"glm-4.6","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":" the"}}]}

But when I try to send request from opencode it finishes by timeout after long waiting.

debug log print looks like:

INFO  2025-11-15T09:07:55 +169ms service=config path=/home/igor/.config/opencode/config.json loading
INFO  2025-11-15T09:07:55 +1ms service=config path=/home/igor/.config/opencode/opencode.json loading
INFO  2025-11-15T09:07:55 +35ms service=config path=/home/igor/.config/opencode/opencode.jsonc loading
INFO  2025-11-15T09:07:55 +51ms service=server method=GET path=/config/providers request
INFO  2025-11-15T09:07:55 +0ms service=server status=started method=GET path=/config/providers request
INFO  2025-11-15T09:07:55 +1ms service=default directory=/home/igor/Plastov/lightconsole_soft/lightconsole_test_eobj creating instance
INFO  2025-11-15T09:07:55 +1ms service=project directory=/home/igor/Plastov/lightconsole_soft/lightconsole_test_eobj fromDirectory
INFO  2025-11-15T09:07:55 +0ms service=server method=GET path=/agent request
INFO  2025-11-15T09:07:55 +0ms service=server status=started method=GET path=/agent request
INFO  2025-11-15T09:07:55 +1ms service=server method=GET path=/config request
INFO  2025-11-15T09:07:55 +0ms service=server status=started method=GET path=/config request
INFO  2025-11-15T09:07:55 +0ms service=project status=started git.rev-parse
INFO  2025-11-15T09:07:55 +2ms service=project status=completed duration=2 git.rev-parse
INFO  2025-11-15T09:07:55 +3ms service=server method=GET path=/event request
INFO  2025-11-15T09:07:55 +0ms service=server status=started method=GET path=/event request
INFO  2025-11-15T09:07:55 +4ms service=default directory=/home/igor/Plastov/lightconsole_soft/lightconsole_test_eobj bootstrapping
INFO  2025-11-15T09:07:55 +5ms service=bun cmd=["/home/igor/.opencode/bin/opencode","add","@opencode-ai/plugin@1.0.65","--exact"] cwd=/home/igor/.config/opencode running
INFO  2025-11-15T09:07:55 +7ms service=bun code=0 stdout=bun add v1.3.2 (b131639c)

installed @opencode-ai/plugin@1.0.65

[3.00ms] done
 stderr=Saved lockfile
 done
INFO  2025-11-15T09:07:55 +6ms service=plugin path=opencode-copilot-auth@0.0.5 loading plugin
INFO  2025-11-15T09:07:55 +2ms service=plugin path=opencode-anthropic-auth@0.0.2 loading plugin
INFO  2025-11-15T09:07:55 +30ms service=bus type=* subscribing
INFO  2025-11-15T09:07:55 +0ms service=bus type=session.updated subscribing
INFO  2025-11-15T09:07:55 +0ms service=bus type=message.updated subscribing
INFO  2025-11-15T09:07:55 +0ms service=bus type=message.part.updated subscribing
INFO  2025-11-15T09:07:55 +1ms service=format init
INFO  2025-11-15T09:07:55 +0ms service=bus type=file.edited subscribing
INFO  2025-11-15T09:07:55 +0ms service=lsp serverIds=deno, typescript, vue, eslint, gopls, ruby-lsp, pyright, elixir-ls, zls, csharp, sourcekit-lsp, rust, clangd, svelte, astro, jdtls, lua-ls enabled LSP servers
INFO  2025-11-15T09:07:55 +2ms service=bus type=command.executed subscribing
INFO  2025-11-15T09:07:55 +2ms service=server status=started providers
INFO  2025-11-15T09:07:55 +3ms service=provider status=started state
INFO  2025-11-15T09:07:55 +1ms service=server event connected
INFO  2025-11-15T09:07:55 +3ms service=bus type=* subscribing
INFO  2025-11-15T09:07:55 +3ms service=models.dev file={} refreshing
INFO  2025-11-15T09:07:55 +3ms service=server status=completed duration=72 method=GET path=/event request
INFO  2025-11-15T09:07:55 +0ms service=server status=completed duration=77 method=GET path=/config request
INFO  2025-11-15T09:07:55 +1ms service=server status=completed duration=79 method=GET path=/agent request
INFO  2025-11-15T09:07:55 +8ms service=provider init
INFO  2025-11-15T09:07:55 +16ms service=provider providerID=zai-coding-plan found
INFO  2025-11-15T09:07:55 +0ms service=provider providerID=opencode found
INFO  2025-11-15T09:07:55 +0ms service=provider status=completed duration=35 state
INFO  2025-11-15T09:07:55 +2ms service=server status=completed duration=40 providers
INFO  2025-11-15T09:07:55 +0ms service=server status=completed duration=107 method=GET path=/config/providers request
INFO  2025-11-15T09:07:55 +16ms service=server method=GET path=/session request
INFO  2025-11-15T09:07:55 +0ms service=server status=started method=GET path=/session request
INFO  2025-11-15T09:07:55 +1ms service=server method=GET path=/command request
INFO  2025-11-15T09:07:55 +0ms service=server status=started method=GET path=/command request
INFO  2025-11-15T09:07:55 +1ms service=server status=completed duration=1 method=GET path=/command request
INFO  2025-11-15T09:07:55 +1ms service=server method=GET path=/lsp request
INFO  2025-11-15T09:07:55 +0ms service=server status=started method=GET path=/lsp request
INFO  2025-11-15T09:07:55 +0ms service=server status=completed duration=0 method=GET path=/lsp request
INFO  2025-11-15T09:07:55 +1ms service=server method=GET path=/mcp request
INFO  2025-11-15T09:07:55 +1ms service=server status=started method=GET path=/mcp request
INFO  2025-11-15T09:07:55 +0ms service=server status=completed duration=1 method=GET path=/mcp request
INFO  2025-11-15T09:07:55 +0ms service=server method=GET path=/formatter request
INFO  2025-11-15T09:07:55 +0ms service=server status=started method=GET path=/formatter request
INFO  2025-11-15T09:07:55 +8ms service=server status=completed duration=13 method=GET path=/session request
INFO  2025-11-15T09:07:55 +1ms service=server status=completed duration=9 method=GET path=/formatter request
ERROR 2025-11-15T09:08:05 +9942ms service=default e=The operation timed out. rejection
INFO  2025-11-15T09:08:09 +3551ms service=server method=POST path=/session request
INFO  2025-11-15T09:08:09 +0ms service=server status=started method=POST path=/session request
INFO  2025-11-15T09:08:09 +4ms service=session id=ses_5793a61e1ffeJII1qOvBi9wTIy version=1.0.65 projectID=c98cea3e79beb9bb24002455891052cf3c497e43 directory=/home/igor/Plastov/lightconsole_soft/lightconsole_test_eobj title=New session - 2025-11-15T09:08:09.374Z time={"created":1763197689374,"updated":1763197689374} created
INFO  2025-11-15T09:08:09 +1ms service=bus type=session.created publishing
INFO  2025-11-15T09:08:09 +1ms service=bus type=session.updated publishing
INFO  2025-11-15T09:08:09 +2ms service=server status=completed duration=8 method=POST path=/session request
INFO  2025-11-15T09:08:09 +4ms service=server method=POST path=/session/ses_5793a61e1ffeJII1qOvBi9wTIy/message request
INFO  2025-11-15T09:08:09 +0ms service=server status=started method=POST path=/session/ses_5793a61e1ffeJII1qOvBi9wTIy/message request
INFO  2025-11-15T09:08:09 +12ms service=session.prompt session=ses_5793a61e1ffeJII1qOvBi9wTIy prompt
INFO  2025-11-15T09:08:09 +0ms service=server status=completed duration=12 method=POST path=/session/ses_5793a61e1ffeJII1qOvBi9wTIy/message request
INFO  2025-11-15T09:08:09 +8ms service=bus type=message.updated publishing
INFO  2025-11-15T09:08:09 +6ms service=bus type=message.part.updated publishing
INFO  2025-11-15T09:08:09 +2ms service=bus type=session.updated publishing
INFO  2025-11-15T09:08:09 +1ms service=provider providerID=zai-coding-plan modelID=glm-4.6 getModel
INFO  2025-11-15T09:08:09 +1ms service=provider status=started providerID=zai-coding-plan getSDK
INFO  2025-11-15T09:08:09 +127ms service=provider status=completed duration=127 providerID=zai-coding-plan getSDK
INFO  2025-11-15T09:08:09 +3ms service=provider providerID=zai-coding-plan modelID=glm-4.6 found
INFO  2025-11-15T09:08:09 +1ms service=session.lock sessionID=ses_5793a61e1ffeJII1qOvBi9wTIy locked
INFO  2025-11-15T09:08:09 +0ms service=session.prompt session=ses_5793a61e1ffeJII1qOvBi9wTIy sessionID=ses_5793a61e1ffeJII1qOvBi9wTIy locking
INFO  2025-11-15T09:08:09 +65ms service=bus type=message.updated publishing
INFO  2025-11-15T09:08:09 +14ms service=session.prompt session=ses_5793a61e1ffeJII1qOvBi9wTIy process
INFO  2025-11-15T09:08:09 +21ms service=bus type=message.updated publishing
INFO  2025-11-15T09:08:09 +1ms service=bus type=session.updated publishing
INFO  2025-11-15T09:08:09 +0ms service=bus type=session.diff publishing
INFO  2025-11-15T09:08:11 +1830ms service=bus type=session.updated publishing
ERROR 2025-11-15T09:12:55 +284229ms service=session.prompt session=ses_5793a61e1ffeJII1qOvBi9wTIy error={"error":{}} stream error
ERROR 2025-11-15T09:12:55 +3ms service=session.prompt session=ses_5793a61e1ffeJII1qOvBi9wTIy error=The operation timed out. process
INFO  2025-11-15T09:12:55 +0ms service=bus type=session.error publishing
INFO  2025-11-15T09:12:55 +4ms service=bus type=message.updated publishing
INFO  2025-11-15T09:12:55 +1ms service=bus type=message.updated publishing
INFO  2025-11-15T09:12:55 +2ms service=session.compaction pruning
INFO  2025-11-15T09:12:55 +1ms service=session.lock sessionID=ses_5793a61e1ffeJII1qOvBi9wTIy unlocked
INFO  2025-11-15T09:12:55 +0ms service=session.prompt session=ses_5793a61e1ffeJII1qOvBi9wTIy sessionID=ses_5793a61e1ffeJII1qOvBi9wTIy unlocking
INFO  2025-11-15T09:12:55 +2ms service=bus type=session.idle publishing
INFO  2025-11-15T09:12:55 +2ms service=bus type=* unsubscribing
INFO  2025-11-15T09:12:55 +0ms service=server event disconnected
INFO  2025-11-15T09:12:55 +1ms service=session.compaction pruned=0 total=0 found
INFO  2025-11-15T09:12:58 +2987ms service=server method=GET path=/event request
INFO  2025-11-15T09:12:58 +1ms service=server status=started method=GET path=/event request
INFO  2025-11-15T09:12:58 +0ms service=server event connected
INFO  2025-11-15T09:12:58 +0ms service=bus type=* subscribing
INFO  2025-11-15T09:12:58 +1ms service=server status=completed duration=1 method=GET path=/event request
INFO  2025-11-15T09:17:55 +296993ms service=bus type=* unsubscribing
INFO  2025-11-15T09:17:55 +0ms service=server event disconnected
INFO  2025-11-15T09:18:01 +6003ms service=server method=GET path=/event request
INFO  2025-11-15T09:18:01 +1ms service=server status=started method=GET path=/event request
INFO  2025-11-15T09:18:01 +0ms service=server event connected
INFO  2025-11-15T09:18:01 +1ms service=bus type=* subscribing
INFO  2025-11-15T09:18:01 +0ms service=server status=completed duration=1 method=GET path=/event request
INFO  2025-11-15T09:21:28 +207261ms service=bus type=* unsubscribing
INFO  2025-11-15T09:21:28 +0ms service=server event disconnected
INFO  2025-11-15T09:21:28 +1ms service=default worker shutting down
INFO  2025-11-15T09:21:28 +0ms service=default disposing all instances
INFO  2025-11-15T09:21:28 +1ms service=default directory=/home/igor/Plastov/lightconsole_soft/lightconsole_test_eobj disposing instance
INFO  2025-11-15T09:21:28 +1ms service=state key=/home/igor/Plastov/lightconsole_soft/lightconsole_test_eobj waiting for state disposal to complete
INFO  2025-11-15T09:21:28 +1ms service=state key=/home/igor/Plastov/lightconsole_soft/lightconsole_test_eobj state disposal completed

How to fix this issue?

Best regards,
Igor

[github-actions]

This issue might be a duplicate of existing issues. Please check:

• OpenCode v1.0.25 Hangs Indefinitely with LM Studio + Qwen Models Due to Empty tool_calls Array #4255: Very similar indefinite hanging issue with detailed root cause analysis pointing to tool calling problems
• No Response Requested #2892: GLM 4.6 timeout/hanging issue where users get stuck on generating with no code
• have error with tool call from 1.0.62 build ( current 1.0.65 still have this issue ) #4322: Z.ai Coding Plan GLM 4.6 with tool calling issues since v1.0.62

The core issue appears to be related to improper handling of empty tool_calls: [] arrays that some providers (including Z.ai) include in responses, causing OpenCode to hang indefinitely.

Feel free to ignore if none of these address your specific case.

[chetverovod]

Addition to opencode.json "tools" fieild:

"tools": {
    "bash": false,
    "read": false,
    "write": false,
    "edit": false,
    "glob": false,
    "grep": false,
    "list": false,
    "todoread": false,
    "todowrite": false,
    "webfetch": false,
    "task": false
  }

Fixes connection. But maybe cuts functionality.

[dumheter]

Try using the z.ai coding plan instead, it has its own auth method in opencode.

[rekram1-node]

@chetverovod add this to your opencode.json:

{
 "provider": {
   "zai-coding-plan": {
     "options": {
       "timeout": 600000
     }
   }
 }
}

This will override default timeout for the provider and set it to 10 minutes (timeout is in milliseconds)

[rekram1-node]

The fetch automatically times out if there is no data being sent from the server at all, as long as data is being sent the connection won't timeout

[mossbergmaverick]

Try using the z.ai coding plan instead, it has its own auth method in opencode.

The z.ai coding plan doesn't work with OpenCode in VS Code extension, although it works in the terminal.

[rekram1-node]

@mossbergmaverick hm that doesn't make sense, is it the oauth flow not going through properly? The vscode extension just launches the tui

[mossbergmaverick]

@mossbergmaverick hm that doesn't make sense, is it the oauth flow not going through properly? The vscode extension just launches the tui

I'll describe in more detail what I mean:
I opened VSCode and launched it through the extension (I have Sonnet 4.5 as the main agent and GLM 4.6 as a sub-agent, to which tasks are delegated). I asked the agent to pass a task to the sub-agent to create a test text file in the project root. Nothing happened - the model looks like it froze. However, if I connect GLM directly and interact with it, it responds.
I just rechecked, and the problem is that GLM should request permission for write and edit operations, but the confirmation dialog doesn't appear on my end. I thought the model had simply frozen. But as I understand it now, this is related to my issue - meaning not only does ESC not work, but the confirmation dialog for executing actions is also missing.
If you need me to take any specific steps to clarify the situation, I'm ready to help.

[chetverovod]

@chetverovod add this to your opencode.json:

{
 "provider": {
   "zai-coding-plan": {
     "options": {
       "timeout": 600000
     }
   }
 }
}

This will override default timeout for the provider and set it to 10 minutes (timeout is in milliseconds)

Hello, now for opencode.json:

 {
     "$schema": "https://opencode.ai/config.json",
     "provider": {
         "zai-coding-plan": {
             "options": {
                 "timeout": 600000
             }
         }
     }
 }

I had got timeout after 10 minutes:

[rekram1-node]

@chetverovod you can continue bumping the timeout to higher values, we can consider changing the default too I suppose

[rekram1-node]

oh u can disable the timeout too using

timeout: false

[rekram1-node]

@mossbergmaverick ill check it out