github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #10899] qwen2.5-coder and llama3.1 return empty content with tools #69223

New Issue
Closed
opened 2026-05-04 17:30:27 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-05-04 17:30:27 -05:00
Owner
Copy Link

Originally created by @rensftw on GitHub (May 29, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10899

Originally assigned to: @ParthSareen on GitHub.

What is the issue?

Description

When I call a model with tools, I always get an empty response from Ollama (like "content": "")
I have observed this behaviour with:

  • qwen2.5-coder:7b
  • llama3.1:latest
  • devstral:latest

Interestingly, qwen3:8b does work with tools and correctly streams back a response.

Example call:

Request:

curl http://localhost:11434/api/chat -d '{
  "model": "qwen2.5-coder:7b",
  "messages": [
    {
      "role": "user",
      "content": "What is the weather today in Toronto?"
    }
  ],
  "stream": true,
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "get_current_weather",
        "description": "Get the current weather for a location",
        "parameters": {
          "type": "object",
          "properties": {
            "location": {
              "type": "string",
              "description": "The location to get the weather for, e.g. San Francisco, CA"
            },
            "format": {
              "type": "string",
              "description": "The format to return the weather in, e.g. 'celsius' or 'fahrenheit'",
              "enum": ["celsius", "fahrenheit"]
            }
          },
          "required": ["location", "format"]
        }
      }
    }
  ]
}'

Response:

{
  "model": "qwen2.5-coder:7b",
  "created_at": "2025-05-29T15:32:39.778126Z",
  "message": {
    "role": "assistant",
    "content": "",
    "tool_calls": [
      {
        "function": {
          "name": "get_current_weather",
          "arguments": {
            "format": "celsius",
            "location": "Toronto"
          }
        }
      }
    ]
  },
  "done": false
}
{
  "model": "qwen2.5-coder:7b",
  "created_at": "2025-05-29T15:32:39.815178Z",
  "message": {
    "role": "assistant",
    "content": ""
  },
  "done_reason": "stop",
  "done": true,
  "total_duration": 2543474417,
  "load_duration": 676834417,
  "prompt_eval_count": 224,
  "prompt_eval_duration": 972060750,
  "eval_count": 25,
  "eval_duration": 890930417
}

Relevant log output

OS

macOS

GPU

Apple

CPU

Apple

Ollama version

0.8.0

Originally created by @rensftw on GitHub (May 29, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10899 Originally assigned to: @ParthSareen on GitHub. ### What is the issue? # Description When I call a model with tools, I **always** get an empty response from Ollama (like `"content": ""`) I have observed this behaviour with: - `qwen2.5-coder:7b` - `llama3.1:latest` - `devstral:latest` Interestingly, `qwen3:8b` does work with tools and correctly streams back a response. ## Example call: Request: ```sh curl http://localhost:11434/api/chat -d '{ "model": "qwen2.5-coder:7b", "messages": [ { "role": "user", "content": "What is the weather today in Toronto?" } ], "stream": true, "tools": [ { "type": "function", "function": { "name": "get_current_weather", "description": "Get the current weather for a location", "parameters": { "type": "object", "properties": { "location": { "type": "string", "description": "The location to get the weather for, e.g. San Francisco, CA" }, "format": { "type": "string", "description": "The format to return the weather in, e.g. 'celsius' or 'fahrenheit'", "enum": ["celsius", "fahrenheit"] } }, "required": ["location", "format"] } } } ] }' ``` Response: ```json { "model": "qwen2.5-coder:7b", "created_at": "2025-05-29T15:32:39.778126Z", "message": { "role": "assistant", "content": "", "tool_calls": [ { "function": { "name": "get_current_weather", "arguments": { "format": "celsius", "location": "Toronto" } } } ] }, "done": false } { "model": "qwen2.5-coder:7b", "created_at": "2025-05-29T15:32:39.815178Z", "message": { "role": "assistant", "content": "" }, "done_reason": "stop", "done": true, "total_duration": 2543474417, "load_duration": 676834417, "prompt_eval_count": 224, "prompt_eval_duration": 972060750, "eval_count": 25, "eval_duration": 890930417 } ``` ### Relevant log output ```shell ``` ### OS macOS ### GPU Apple ### CPU Apple ### Ollama version 0.8.0
GiteaMirror added the toolsbug labels 2026-05-04 17:30:28 -05:00
GiteaMirror commented 2026-05-04 17:30:42 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (May 29, 2025):

For most non-thinking tools, the content is the tool call. The context gets converted into the tool_calls structure and is cleared.

5f57b0ef42/server/routes.go (L1581)

qwen3:8b, as a thinking model, has the tool call removed from the output but the remaining thinking is returned.

<!-- gh-comment-id:2919867957 --> @rick-github commented on GitHub (May 29, 2025): For most non-thinking tools, the content is the tool call. The context gets converted into the `tool_calls` structure and is cleared. https://github.com/ollama/ollama/blob/5f57b0ef4268a6bd9e8043d54c351a608a7e1bca/server/routes.go#L1581 qwen3:8b, as a thinking model, has the tool call removed from the output but the remaining thinking is returned.
GiteaMirror commented 2026-05-04 17:30:43 -05:00
Author
Owner
Copy Link

@rensftw commented on GitHub (May 29, 2025):

@rick-github You're right

I hadn't read all of the chunked responses until now - looks like qwen3:8b prints only the thinking part and not the actual result from the tool.

<!-- gh-comment-id:2919932481 --> @rensftw commented on GitHub (May 29, 2025): @rick-github You're right I hadn't read all of the chunked responses until now - looks like `qwen3:8b` prints only the thinking part and not the actual result from the tool.
GiteaMirror commented 2026-05-04 17:30:46 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (May 29, 2025):

Note that in 0.9.0 this will be returned in the thinking field and content will be empty again.

<!-- gh-comment-id:2919936483 --> @rick-github commented on GitHub (May 29, 2025): Note that in 0.9.0 this will be returned in the `thinking` field and `content` will be empty again.
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 tools
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#69223
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?