[GH-ISSUE #15690] issue: tool calls fail when model makes multiple tool calls in one response #33169

New Issue

GiteaMirror · 2026-04-25T07:04:16-05:00

GiteaMirror commented

2026-04-25 07:04:16 -05:00

Originally created by @Master-Pr0grammer on GitHub (Jul 13, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/15690

Check Existing Issues

I have searched the existing issues and discussions.
I am using the latest version of Open WebUI.

Installation Method

Pip Install

Open WebUI Version

v0.6.15

Ollama Version (if applicable)

0.9.6

Operating System

Ubuntu 24.04.1

Browser (if applicable)

Chrome

Confirmation

I have read and followed all instructions in README.md.
I am using the latest version of both Open WebUI and Ollama.
I have included the browser console logs.
I have included the Docker container logs.
I have provided every relevant configuration, setting, and environment variable used in my setup.
I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
Start with the initial platform/version/OS and dependencies used,
Specify exact install/launch/configure commands,
List URLs visited, user input (incl. example values/emails/passwords if needed),
Describe all options and toggles enabled or changed,
Include any files or environmental changes,
Identify the expected and actual result at each stage,
Ensure any reasonably skilled user can follow and hit the same issue.

Expected Behavior

when the model calls two tools, they should be both be executed individually (in order).

Actual Behavior

openwebui gets confused and just ends the chat completion run.

for example if the tool is print(str), and it gets called twice, openwebui will output Executing printprint..., and immediately end there without executing or continuing to generate the next response

Steps to Reproduce

get a model that is capable of outputting multiple tool calls (I used qwen3:8b)
give it access to a tool
ask it to use the tool twice in a row to do two independent tasks

Logs & Screenshots

index.js:1183 [tiptap warn]: Duplicate extension names found: ['codeBlock']. This can lead to issues.
resolve @ index.js:1183
ResponseMessage.svelte:573 [Violation] Added non-passive event listener to a scroll-blocking 'wheel' event. Consider marking event handler as 'passive' to make the page more responsive. See https://www.chromestatus.com/feature/5745543795965952
(anonymous) @ ResponseMessage.svelte:573

Additional Information

This is not a model problem, i have tried several models, and all of them had the same issue in openwebui. when i run the same prompt through my own ollama api based script, it has no issues calling the tools.

Originally created by @Master-Pr0grammer on GitHub (Jul 13, 2025). Original GitHub issue: https://github.com/open-webui/open-webui/issues/15690 ### Check Existing Issues - [x] I have searched the existing issues and discussions. - [x] I am using the latest version of Open WebUI. ### Installation Method Pip Install ### Open WebUI Version v0.6.15 ### Ollama Version (if applicable) 0.9.6 ### Operating System Ubuntu 24.04.1 ### Browser (if applicable) Chrome ### Confirmation - [x] I have read and followed all instructions in `README.md`. - [x] I am using the latest version of **both** Open WebUI and Ollama. - [x] I have included the browser console logs. - [x] I have included the Docker container logs. - [x] I have **provided every relevant configuration, setting, and environment variable used in my setup.** - [x] I have clearly **listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup** (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc). - [x] I have documented **step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation**. My steps: - Start with the initial platform/version/OS and dependencies used, - Specify exact install/launch/configure commands, - List URLs visited, user input (incl. example values/emails/passwords if needed), - Describe all options and toggles enabled or changed, - Include any files or environmental changes, - Identify the expected and actual result at each stage, - Ensure any reasonably skilled user can follow and hit the same issue. ### Expected Behavior when the model calls two tools, they should be both be executed individually (in order). ### Actual Behavior openwebui gets confused and just ends the chat completion run. for example if the tool is print(str), and it gets called twice, openwebui will output `Executing printprint...`, and immediately end there without executing or continuing to generate the next response ### Steps to Reproduce 1. get a model that is capable of outputting multiple tool calls (I used qwen3:8b) 2. give it access to a tool 3. ask it to use the tool twice in a row to do two independent tasks ### Logs & Screenshots index.js:1183 [tiptap warn]: Duplicate extension names found: ['codeBlock']. This can lead to issues. resolve @ index.js:1183 ResponseMessage.svelte:573 [Violation] Added non-passive event listener to a scroll-blocking 'wheel' event. Consider marking event handler as 'passive' to make the page more responsive. See https://www.chromestatus.com/feature/5745543795965952 (anonymous) @ ResponseMessage.svelte:573 <img width="1203" height="792" alt="Image" src="https://github.com/user-attachments/assets/b2113d6e-d87c-4102-957f-4e6f6c365579" /> ### Additional Information This is not a model problem, i have tried several models, and all of them had the same issue in openwebui. when i run the same prompt through my own ollama api based script, it has no issues calling the tools.

GiteaMirror added the bug label 2026-04-25 07:04:16 -05:00

GiteaMirror closed this issue

2026-04-25 07:04:17 -05:00

GiteaMirror commented

2026-04-25 07:04:17 -05:00

@GlisseManTV commented on GitHub (Jul 13, 2025):

Hi multiples posts are already answered yet.
It's about an tool ID/name in code.
You could check this :
PR14680
I13726
I14583
and more
Empty field is from 0.6.15 and after.
on 0.6.14 you'll have wrong tool call.

@GlisseManTV commented on GitHub (Jul 13, 2025): Hi multiples posts are already answered yet. It's about an tool ID/name in code. You could check this : [PR14680](https://github.com/open-webui/open-webui/pull/14680) [I13726](https://github.com/open-webui/open-webui/issues/13726) [I14583](https://github.com/open-webui/open-webui/issues/14583) and more Empty field is from 0.6.15 and after. on 0.6.14 you'll have wrong tool call.

GiteaMirror commented

2026-04-25 07:04:17 -05:00

@tjbck commented on GitHub (Jul 14, 2025):

Could you confirm if this is still the case with 0.6.16? We were unable to reproduce with higher end models like gpt-4.1. Keep us updated!

@tjbck commented on GitHub (Jul 14, 2025): Could you confirm if this is still the case with 0.6.16? We were unable to reproduce with higher end models like `gpt-4.1`. Keep us updated!

Sign in to join this conversation.

Branches Tags

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/open-webui#33169