GLM Coding Plan - Frequent Errors #1905

Closed
opened 2026-02-16 17:33:10 -05:00 by yindo · 20 comments
Owner

Originally created by @cvkmohan on GitHub (Oct 1, 2025).

Originally assigned to: @thdxr on GitHub.

With GLM I am getting a regular "Invalid API Parameter. Check the documentation." as well as a timeout error. Both are frequent - and - I have to type "continue" to resume the activity. Not getting similar errors for other models. Should I be increasing some time interval somewhere? Just want to bring to your notice. This is more pronounced in 4.6 rather than 4.5

Originally created by @cvkmohan on GitHub (Oct 1, 2025). Originally assigned to: @thdxr on GitHub. With GLM I am getting a regular "Invalid API Parameter. Check the documentation." as well as a timeout error. Both are frequent - and - I have to type "continue" to resume the activity. Not getting similar errors for other models. Should I be increasing some time interval somewhere? Just want to bring to your notice. This is more pronounced in 4.6 rather than 4.5
yindo closed this issue 2026-02-16 17:33:10 -05:00
Author
Owner

@github-actions[bot] commented on GitHub (Oct 1, 2025):

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

  • #2557: Similar GLM-4.5 API parameter errors with "The messages parameter is illegal. Please check the documentation." and requiring "continue" to resume
  • #1880: GLM 4.5 string/response data errors that cause opencode to stop working
  • #2892: GLM 4.6 "No Response Requested" errors and hanging issues
  • #2774: General timeout issues requiring "continue" to resume activity

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

@github-actions[bot] commented on GitHub (Oct 1, 2025): This issue might be a duplicate of existing issues. Please check: - #2557: Similar GLM-4.5 API parameter errors with "The messages parameter is illegal. Please check the documentation." and requiring "continue" to resume - #1880: GLM 4.5 string/response data errors that cause opencode to stop working - #2892: GLM 4.6 "No Response Requested" errors and hanging issues - #2774: General timeout issues requiring "continue" to resume activity Feel free to ignore if none of these address your specific case.
Author
Owner

@rekram1-node commented on GitHub (Oct 1, 2025):

Hm this thread may be helpful (regarding api parameter issues): https://github.com/sst/opencode/issues/2557

You can increase timeouts for a provider if you want:

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

But generally I think the timeout is just the providers fault, if I notice requests hanging I like to double ESC then type "." and hit enter

@rekram1-node commented on GitHub (Oct 1, 2025): Hm this thread may be helpful (regarding api parameter issues): https://github.com/sst/opencode/issues/2557 You can increase timeouts for a provider if you want: ``` { "$schema": "https://opencode.ai/config.json", "provider": { "zai-coding-plan": { "options": { "timeout": 600000 } } } } ``` But generally I think the timeout is just the providers fault, if I notice requests hanging I like to double ESC then type "." and hit enter
Author
Owner

@ShamanicArts commented on GitHub (Oct 2, 2025):

yeah, I have just started getting tons of AI_APICallError: Invalid API parameter, please check the documentation. errors on GLM as well now, not sure the cause.

@ShamanicArts commented on GitHub (Oct 2, 2025): yeah, I have just started getting tons of ```AI_APICallError: Invalid API parameter, please check the documentation.``` errors on GLM as well now, not sure the cause.
Author
Owner

@rekram1-node commented on GitHub (Oct 2, 2025):

@ShamanicArts which provider?

@rekram1-node commented on GitHub (Oct 2, 2025): @ShamanicArts which provider?
Author
Owner

@ShamanicArts commented on GitHub (Oct 2, 2025):

Z.AI Coding Plan

@ShamanicArts commented on GitHub (Oct 2, 2025): Z.AI Coding Plan
Author
Owner

@ShamanicArts commented on GitHub (Oct 2, 2025):

https://opencode.ai/s/X4oarVXT @rekram1-node

@ShamanicArts commented on GitHub (Oct 2, 2025): https://opencode.ai/s/X4oarVXT @rekram1-node
Author
Owner

@Rybens92 commented on GitHub (Oct 2, 2025):

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

Zai-coding-plan seams to have problems with heavy traffic, I think, so editing provider ~/.config/opencode/opencode.json with this change works (at least for me).

@Rybens92 commented on GitHub (Oct 2, 2025): > "zai-coding-plan": { > "options": { > "timeout": 600000 > } > } Zai-coding-plan seams to have problems with heavy traffic, I think, so editing provider ~/.config/opencode/opencode.json with this change works (at least for me).
Author
Owner

@rekram1-node commented on GitHub (Oct 2, 2025):

Thx for calling this out @Rybens92

Note: I do think Zai is having a lot of traffic and scale issues, I heard several people report outages today not sure if it relates or not

@rekram1-node commented on GitHub (Oct 2, 2025): Thx for calling this out @Rybens92 Note: I do think Zai is having a lot of traffic and scale issues, I heard several people report outages today not sure if it relates or not
Author
Owner

@Longhuiberkeley commented on GitHub (Oct 2, 2025):

I tried that. Yeah, I think it's just traffic issue. For people who are also experiencing it, try what @rekram1-node suggested. Like do a nano ~/.config/opencode/opencode.json (save you a click, if you don't know where the config file goes), and paste that.

I am still having the error even with the timeout config. Just send a empty message . like @rekram1-node is good advice. Thanks for your messages here

@Longhuiberkeley commented on GitHub (Oct 2, 2025): I tried that. Yeah, I think it's just traffic issue. For people who are also experiencing it, try what @rekram1-node suggested. Like do a `nano ~/.config/opencode/opencode.json` (save you a click, if you don't know where the config file goes), and paste that. I am still having the error even with the timeout config. Just send a empty message `.` like @rekram1-node is good advice. Thanks for your messages here
Author
Owner

@rekram1-node commented on GitHub (Oct 2, 2025):

Happy to help! any issues you have just lmk

@rekram1-node commented on GitHub (Oct 2, 2025): Happy to help! any issues you have just lmk
Author
Owner

@ShamanicArts commented on GitHub (Oct 2, 2025):

yes, fully resolved for me, think it was just outage symptoms!

@ShamanicArts commented on GitHub (Oct 2, 2025): yes, fully resolved for me, think it was just outage symptoms!
Author
Owner

@rekram1-node commented on GitHub (Oct 2, 2025):

sweet

@rekram1-node commented on GitHub (Oct 2, 2025): sweet
Author
Owner

@Longhuiberkeley commented on GitHub (Oct 3, 2025):

For people who are having the issue still, I think setting timeout helps, but the issue still persists. Likely the ZAI's side is experiencing heavy traffic

@Longhuiberkeley commented on GitHub (Oct 3, 2025): For people who are having the issue still, I think setting timeout helps, but the issue still persists. Likely the ZAI's side is experiencing heavy traffic
Author
Owner

@rekram1-node commented on GitHub (Oct 3, 2025):

Hmm it if keeps happening we may wanna reopen this, this weekend I will get the subscription so I can start playing with it more and maybe encounter the issue

@rekram1-node commented on GitHub (Oct 3, 2025): Hmm it if keeps happening we may wanna reopen this, this weekend I will get the subscription so I can start playing with it more and maybe encounter the issue
Author
Owner

@kierr commented on GitHub (Oct 6, 2025):

I have been running opencode with extreme usage via ZAI Coding Plan with glm-4.6 in the past few days, have not encountered this issue once.

I have had transient issues: 50x and rate limit issues (which 'Continue' resolves, or failing that starting a new session). #2892 is worst of all, because there seems to be no way of recovering.

Coming from a claude code perspective, it might help to have a retry configuration option on API failures?

@kierr commented on GitHub (Oct 6, 2025): I have been running opencode with extreme usage via ZAI Coding Plan with glm-4.6 in the past few days, have not encountered this issue once. I have had transient issues: 50x and rate limit issues (which 'Continue' resolves, or failing that starting a new session). #2892 is worst of all, because there seems to be no way of recovering. Coming from a claude code perspective, it might help to have a retry configuration option on API failures?
Author
Owner

@rekram1-node commented on GitHub (Oct 6, 2025):

Yeah we do have retries currently, need to figure out what is going on with 2892 tho

@rekram1-node commented on GitHub (Oct 6, 2025): Yeah we do have retries currently, need to figure out what is going on with 2892 tho
Author
Owner

@Longhuiberkeley commented on GitHub (Oct 6, 2025):

Reporting back, since I wrote that message 3 days ago. I haven't seen the issue again. So that's good news for me as a user 😀

@Longhuiberkeley commented on GitHub (Oct 6, 2025): Reporting back, since I wrote that message 3 days ago. I haven't seen the issue again. So that's good news for me as a user 😀
Author
Owner

@felipematos commented on GitHub (Oct 13, 2025):

Hi! I have frequent timeout errors using GLM models. Is there any way to solve this?

@felipematos commented on GitHub (Oct 13, 2025): Hi! I have frequent timeout errors using GLM models. Is there any way to solve this?
Author
Owner

@alin231127 commented on GitHub (Nov 27, 2025):

I have the same issue whenever i try to use the screenshot tool for playright mcp in opencode. I tried the timeout options suggested here that dod not do the trick... maybe should i clear opencode cache ? ...i have GLM-4.6 coding plan pro ... i did not go for max ... shouod i upgrade ... what is your experience with this issue ?

@alin231127 commented on GitHub (Nov 27, 2025): I have the same issue whenever i try to use the screenshot tool for playright mcp in opencode. I tried the timeout options suggested here that dod not do the trick... maybe should i clear opencode cache ? ...i have GLM-4.6 coding plan pro ... i did not go for max ... shouod i upgrade ... what is your experience with this issue ?
Author
Owner

@rekram1-node commented on GitHub (Nov 27, 2025):

@alin231127 what version r u on

@rekram1-node commented on GitHub (Nov 27, 2025): @alin231127 what version r u on
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: anomalyco/opencode#1905