github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-16 05:05:37 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #11155] Mistral-small-3.2:24b - tool calling issues #7358

New Issue
Closed
opened 2026-04-12 19:24:47 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-12 19:24:47 -05:00
Owner
Copy Link

Originally created by @Notbici on GitHub (Jun 21, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11155

What is the issue?

Hey,

Native tool calling from Mistral 3.2 (2506) 24B (FP16) seems to be throwing this error at Ollama: "failed to create tool parser"

I've seen in other places that it could be a bad tool call, not sure how to debug firther. I tried the Q4, Q8 and FP16 to rule out weirdness in quantizationsw

Relevant log output

OS

No response

GPU

No response

CPU

No response

Ollama version

0.9.0 (the latest version has pending issues)

Originally created by @Notbici on GitHub (Jun 21, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11155 ### What is the issue? Hey, Native tool calling from Mistral 3.2 (2506) 24B (FP16) seems to be throwing this error at Ollama: "failed to create tool parser" I've seen in other places that it could be a bad tool call, not sure how to debug firther. I tried the Q4, Q8 and FP16 to rule out weirdness in quantizationsw ### Relevant log output ```shell ``` ### OS _No response_ ### GPU _No response_ ### CPU _No response_ ### Ollama version 0.9.0 (the latest version has pending issues)
GiteaMirror added the bug label 2026-04-12 19:24:47 -05:00
GiteaMirror commented 2026-04-12 19:24:49 -05:00
Author
Owner
Copy Link

@Notbici commented on GitHub (Jun 21, 2025):

Could it be due to them making changes to the model in order to be more robust?
535b4d0a0f/src/mistral_common/tokens/tokenizers/instruct.py (L778)

https://huggingface.co/mistralai/Mistral-Small-3.2-24B-Instruct-2506#function-calling

<!-- gh-comment-id:2993565767 --> @Notbici commented on GitHub (Jun 21, 2025): Could it be due to them making changes to the model in order to be more robust? https://github.com/mistralai/mistral-common/blob/535b4d0a0fc94674ea17db6cf8dc2079b81cbcfa/src/mistral_common/tokens/tokenizers/instruct.py#L778 https://huggingface.co/mistralai/Mistral-Small-3.2-24B-Instruct-2506#function-calling
GiteaMirror commented 2026-04-12 19:24:50 -05:00
Author
Owner
Copy Link

@sam-braun commented on GitHub (Jul 1, 2025):

@Notbici Hello! Were you able to figure out how to make Mistral-Small-3.2 more reliably call tools?

<!-- gh-comment-id:3024056811 --> @sam-braun commented on GitHub (Jul 1, 2025): @Notbici Hello! Were you able to figure out how to make Mistral-Small-3.2 more reliably call tools?
GiteaMirror commented 2026-04-12 19:24:50 -05:00
Author
Owner
Copy Link

@Notbici commented on GitHub (Jul 2, 2025):

Hey @sam-braun I have to be brutally honest about that model, I didn't end up keeping it.

They claimed it reduced repetitiveness but I found conversationally it was way more repetitive, and tool calling logic was not effective with it wanting to use tools at a whim or not wanting to use them at all when telling it to.

I switched to Qwen3:32b with moderate successes and I mixture better trained conversational models to handle what comes after tool calls.

However 0.9.4 ollama apparently addresses some issues with tool-calling, try that and see how it goes?

<!-- gh-comment-id:3029051785 --> @Notbici commented on GitHub (Jul 2, 2025): Hey @sam-braun I have to be brutally honest about that model, I didn't end up keeping it. They claimed it reduced repetitiveness but I found conversationally it was way more repetitive, and tool calling logic was not effective with it wanting to use tools at a whim or not wanting to use them at all when telling it to. I switched to Qwen3:32b with moderate successes and I mixture better trained conversational models to handle what comes after tool calls. However 0.9.4 ollama apparently addresses some issues with tool-calling, try that and see how it goes?
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#7358
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?