github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-24 14:43:39 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #13820] Error using tools with glm-4.7-flash #9052

New Issue
Open
opened 2026-04-12 21:52:48 -05:00 by GiteaMirror · 5 comments
GiteaMirror commented 2026-04-12 21:52:48 -05:00
Owner
Copy Link

Originally created by @rottedfrog on GitHub (Jan 21, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/13820

What is the issue?

when using glm-4.7-flash:latest with claude code, it frequently stops and loses all context - looking at the logs I get the attached error whenever it happens.

Relevant log output

Jan 20 13:27:49 tartarus-lx ollama[4078349]: time=2026-01-20T13:27:49.426Z level=WARN source=glm46.go:89 msg="glm-4.6 tool call parsing failed" error="failed to parse XML: XML syntax error on line 1: element <tool_call> closed by </arg_key>"

OS

Linux

GPU

AMD

CPU

AMD

Ollama version

0.14.3

Originally created by @rottedfrog on GitHub (Jan 21, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/13820 ### What is the issue? when using `glm-4.7-flash:latest` with claude code, it frequently stops and loses all context - looking at the logs I get the attached error whenever it happens. ### Relevant log output ```shell Jan 20 13:27:49 tartarus-lx ollama[4078349]: time=2026-01-20T13:27:49.426Z level=WARN source=glm46.go:89 msg="glm-4.6 tool call parsing failed" error="failed to parse XML: XML syntax error on line 1: element <tool_call> closed by </arg_key>" ``` ### OS Linux ### GPU AMD ### CPU AMD ### Ollama version 0.14.3
GiteaMirror added the bug label 2026-04-12 21:52:48 -05:00
GiteaMirror commented 2026-04-12 21:52:49 -05:00
Author
Owner
Copy Link

@sommarnatt commented on GitHub (Jan 21, 2026):

I think you need to wait for the updated model - if this one is from yesterday, then at least 1 fix has been added since then.

https://www.reddit.com/r/LocalLLaMA/comments/1qih9r8/current_glm47flash_implementation_confirmed_to_be/
https://www.reddit.com/r/unsloth/comments/1qiu5w8/glm47flash_ggufs_updated_now_produces_much_better/

<!-- gh-comment-id:3780308510 --> @sommarnatt commented on GitHub (Jan 21, 2026): I think you need to wait for the updated model - if this one is from yesterday, then at least 1 fix has been added since then. https://www.reddit.com/r/LocalLLaMA/comments/1qih9r8/current_glm47flash_implementation_confirmed_to_be/ https://www.reddit.com/r/unsloth/comments/1qiu5w8/glm47flash_ggufs_updated_now_produces_much_better/
GiteaMirror commented 2026-04-12 21:52:50 -05:00
Author
Owner
Copy Link

@bherw commented on GitHub (Jan 22, 2026):

I'm using the updated model from unsloth, and both opencode and claude code just immediately report a 400 error from ollama that the model "does not support tools" and don't even try to do anything.

<!-- gh-comment-id:3781899908 --> @bherw commented on GitHub (Jan 22, 2026): I'm using the updated model from unsloth, and both opencode and claude code just immediately report a 400 error from ollama that the model "does not support tools" and don't even try to do anything.
GiteaMirror commented 2026-04-12 21:52:51 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 22, 2026):

@bherw HF models generally don't come with the template that ollama uses to detect capabilities. Either use the model from the ollama library, or merge the library Modelfile with the HF GGUF.

<!-- gh-comment-id:3783558948 --> @rick-github commented on GitHub (Jan 22, 2026): @bherw HF models generally don't come with the template that ollama uses to detect capabilities. Either use the model from the ollama library, or merge the library Modelfile with the HF GGUF.
GiteaMirror commented 2026-04-12 21:52:51 -05:00
Author
Owner
Copy Link

@ParthSareen commented on GitHub (Jan 23, 2026):

Hi @bherw - we have our own parsers which work differently than other templates. YMMV for models outside of Ollama.

@rottedfrog what context length are you running with? I'm finding this model needs more context length and shifting does not work well with tools.

<!-- gh-comment-id:3788325225 --> @ParthSareen commented on GitHub (Jan 23, 2026): Hi @bherw - we have our own parsers which work differently than other templates. YMMV for models outside of Ollama. @rottedfrog what context length are you running with? I'm finding this model needs more context length and shifting does not work well with tools.
GiteaMirror commented 2026-04-12 21:52:52 -05:00
Author
Owner
Copy Link

@balki commented on GitHub (Jan 23, 2026):

Tools worked for me with opencode (with 64K and 128K context)

<!-- gh-comment-id:3791931166 --> @balki commented on GitHub (Jan 23, 2026): Tools worked for me with opencode (with 64K and 128K context)
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#9052
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?