github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #12395] when using cloud model ( Ollama Error: Invalid tool usage: mismatch between tool calls and tool results (status code: 400 ) #33993

New Issue
Open
opened 2026-04-22 17:12:42 -05:00 by GiteaMirror · 9 comments
GiteaMirror commented 2026-04-22 17:12:42 -05:00
Owner
Copy Link

Originally created by @gbonobono on GitHub (Sep 24, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/12395

What is the issue?

Hello thanks for the great code

I am not sure if this issue is belong to your code, but
when I using cloud model, there is error like "Ollama Error: Invalid tool usage: mismatch between tool calls and tool results (status code: 400)"

would you please clarify if this issue is valid?
thanks

========================================================================================
╭────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────── 🔧 Available Tools ───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│ ✓ mcp-server-commands.run_command │
╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────── 1/1 tools enabled ───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
╭──────────────────────────────────────╮
│ 🧠 Current model: gpt-oss:120b-cloud │
╰──────────────────────────────────────╯
gpt-oss/thinking/1-tool❯ 현재폴더에 어떤 파일이 있는지 알려줘

╭─ 🔧 Executing Tool mcp-server-commands.run_command ─╮
│ │
│ Arguments: │
│ │
│ │
│ { │
│ "command": "ls -la" │
│ } │
│ │
╰─────────────────────────────────────────────────────╯

🧑‍💻 Human-in-the-Loop Confirmation
Tool to execute: mcp-server-commands.run_command
Arguments:
• command: ls -la

Options:
y/yes - Execute the tool call
n/no - Skip this tool call
disable - Disable HIL confirmations permanently

What would you like to do? (y): y
Tip: Use 'human-in-loop' or 'hil' to disable these confirmations

╭───────── ✅ Tool Response mcp-server-commands.run_command ──────────╮
│ │
│ Arguments: │
│ │
│ │
│ { │
│ "command": "ls -la" │
│ } │
│ │
│ Response: │
│ │
│ │
│ total 20 │
│ drwxrwxr-x 3 ubuntu ubuntu 4096 Sep 24 14:25 . │
│ drwxr-x--- 32 ubuntu ubuntu 4096 Sep 24 14:25 .. │
│ drwxrwxr-x 8 ubuntu ubuntu 4096 Sep 24 10:37 mcp-server-commands │
│ -rw-rw-r-- 1 ubuntu ubuntu 123 Sep 24 10:39 server.json │
│ -rw-rw-r-- 1 ubuntu ubuntu 6 Sep 24 10:51 temp.txt │
│ │
│ │
╰───────────────────────���─────────────────────────────────────────────╯

⠦ working...
╭───────────────────────────────────────────────────────────────────────────────────────────────────╮
│ Ollama Error: Invalid tool usage: mismatch between tool calls and tool results (status code: 400) │
╰───────────────────────────────────────────────────────────────────────────────────────────────────╯

Relevant log output

OS

No response

GPU

No response

CPU

No response

Ollama version

No response

Originally created by @gbonobono on GitHub (Sep 24, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/12395 ### What is the issue? Hello thanks for the great code I am not sure if this issue is belong to your code, but when I using cloud model, there is error like "Ollama Error: Invalid tool usage: mismatch between tool calls and tool results (status code: 400)" would you please clarify if this issue is valid? thanks ======================================================================================== ╭────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────── 🔧 Available Tools ───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮ │ ✓ mcp-server-commands.run_command │ ╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────── 1/1 tools enabled ───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯ ╭──────────────────────────────────────╮ │ 🧠 Current model: gpt-oss:120b-cloud │ ╰──────────────────────────────────────╯ gpt-oss/thinking/1-tool❯ 현재폴더에 어떤 파일이 있는지 알려줘 ╭─ 🔧 Executing Tool mcp-server-commands.run_command ─╮ │ │ │ Arguments: │ │ │ │ │ │ { │ │ "command": "ls -la" │ │ } │ │ │ ╰─────────────────────────────────────────────────────╯ 🧑‍💻 Human-in-the-Loop Confirmation Tool to execute: mcp-server-commands.run_command Arguments: • command: ls -la Options: y/yes - Execute the tool call n/no - Skip this tool call disable - Disable HIL confirmations permanently What would you like to do? (y): y Tip: Use 'human-in-loop' or 'hil' to disable these confirmations ╭───────── ✅ Tool Response mcp-server-commands.run_command ──────────╮ │ │ │ Arguments: │ │ │ │ │ │ { │ │ "command": "ls -la" │ │ } │ │ │ │ Response: │ │ │ │ │ │ total 20 │ │ drwxrwxr-x 3 ubuntu ubuntu 4096 Sep 24 14:25 . │ │ drwxr-x--- 32 ubuntu ubuntu 4096 Sep 24 14:25 .. │ │ drwxrwxr-x 8 ubuntu ubuntu 4096 Sep 24 10:37 mcp-server-commands │ │ -rw-rw-r-- 1 ubuntu ubuntu 123 Sep 24 10:39 server.json │ │ -rw-rw-r-- 1 ubuntu ubuntu 6 Sep 24 10:51 temp.txt │ │ │ │ │ ╰───────────────────────���─────────────────────────────────────────────╯ ⠦ working... ╭───────────────────────────────────────────────────────────────────────────────────────────────────╮ │ Ollama Error: Invalid tool usage: mismatch between tool calls and tool results (status code: 400) │ ╰───────────────────────────────────────────────────────────────────────────────────────────────────╯ ### Relevant log output ```shell ``` ### OS _No response_ ### GPU _No response_ ### CPU _No response_ ### Ollama version _No response_
GiteaMirror added the cloudbug labels 2026-04-22 17:12:42 -05:00
GiteaMirror commented 2026-04-22 17:12:43 -05:00
Author
Owner
Copy Link

@naknak987 commented on GitHub (Sep 24, 2025):

I also receive this error when using qwen3-coder:480b-cloud.

{"error": "Invalid tool usage: mismatch between tool calls and tool results"}

<!-- gh-comment-id:3330694133 --> @naknak987 commented on GitHub (Sep 24, 2025): I also receive this error when using `qwen3-coder:480b-cloud`. > {"error": "Invalid tool usage: mismatch between tool calls and tool results"}
GiteaMirror commented 2026-04-22 17:12:44 -05:00
Author
Owner
Copy Link

@it-nalon commented on GitHub (Sep 27, 2025):

Hi, I had the same issue and discovered I was sending a tool_call that wasn't declared in the tools array.

I was doing this:

{
  "model": "deepseek-v3.1:671b",
  "messages": [
    {
      "role": "assistant",
      "content": "",
      "tool_calls": [
        {
          "function": {
            "name": "other_agent_tool_call",
            "arguments": {
              ...
            }
          }
        }
      ]
    },
    {
      "role": "user",
      "content": "Provide a concise update on the status based on your last message"
    }
  ],
  "stream": true,
  "tools": [] // <- other_agent_tool_call does not exist here
}

I fixed it by sending a transcript of the message instead of a raw json:

{
  "model": "deepseek-v3.1:671b",
  "messages": [
    {
      "role": "user",
      "content": "Last events:\n[assistant] Tool call: { ... }"
    },
    {
      "role": "user",
      "content": "Provide a concise update on the status based on the last events"
    }
  ],
  "stream": true,
  "tools": []
}

Maybe you're having a similar issue?

<!-- gh-comment-id:3341883537 --> @it-nalon commented on GitHub (Sep 27, 2025): Hi, I had the same issue and discovered I was sending a `tool_call` that wasn't declared in the `tools` array. I was doing this: ``` { "model": "deepseek-v3.1:671b", "messages": [ { "role": "assistant", "content": "", "tool_calls": [ { "function": { "name": "other_agent_tool_call", "arguments": { ... } } } ] }, { "role": "user", "content": "Provide a concise update on the status based on your last message" } ], "stream": true, "tools": [] // <- other_agent_tool_call does not exist here } ``` I fixed it by sending a transcript of the message instead of a raw json: ``` { "model": "deepseek-v3.1:671b", "messages": [ { "role": "user", "content": "Last events:\n[assistant] Tool call: { ... }" }, { "role": "user", "content": "Provide a concise update on the status based on the last events" } ], "stream": true, "tools": [] } ``` Maybe you're having a similar issue?
GiteaMirror commented 2026-04-22 17:12:45 -05:00
Author
Owner
Copy Link

@gbonobono commented on GitHub (Sep 28, 2025):

Hi , the deepseek-v3.1:671b was cloud model?

<!-- gh-comment-id:3342179537 --> @gbonobono commented on GitHub (Sep 28, 2025): Hi , the deepseek-v3.1:671b was cloud model?
GiteaMirror commented 2026-04-22 17:12:45 -05:00
Author
Owner
Copy Link

@it-nalon commented on GitHub (Sep 30, 2025):

Yes, I tried both in the local environment and in the cloud one (ollama.com/api), same behavior.

<!-- gh-comment-id:3350160201 --> @it-nalon commented on GitHub (Sep 30, 2025): Yes, I tried both in the local environment and in the cloud one (`ollama.com/api`), same behavior.
GiteaMirror commented 2026-04-22 17:12:46 -05:00
Author
Owner
Copy Link

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

yes, it happens with opencode too

<!-- gh-comment-id:3397208989 --> @matscalia commented on GitHub (Oct 13, 2025): yes, it happens with opencode too
GiteaMirror commented 2026-04-22 17:12:47 -05:00
Author
Owner
Copy Link

@Emmanuel-Odebiyi commented on GitHub (Oct 14, 2025):

@gbonobono I also had same issue but I found out that it is not all cloud model of ollama actually supports tool calling, so I changed my former model ( gpt-oss:20b-cloud) into another one (gpt-oss:120b-cloud) and that solves the issue, I hope this helps.

<!-- gh-comment-id:3401027986 --> @Emmanuel-Odebiyi commented on GitHub (Oct 14, 2025): @gbonobono I also had same issue but I found out that it is not all cloud model of ollama actually supports tool calling, so I changed my former model ( gpt-oss:20b-cloud) into another one (gpt-oss:120b-cloud) and that solves the issue, I hope this helps.
GiteaMirror commented 2026-04-22 17:12:47 -05:00
Author
Owner
Copy Link

@bitbytebit1 commented on GitHub (Oct 15, 2025):

For me the issue was not pushing the call request onto the message history, this immediately resolved the issue

<!-- gh-comment-id:3407980796 --> @bitbytebit1 commented on GitHub (Oct 15, 2025): For me the issue was not pushing the call request onto the message history, this immediately resolved the issue
GiteaMirror commented 2026-04-22 17:12:48 -05:00
Author
Owner
Copy Link

@jonigl commented on GitHub (Oct 20, 2025):

Hi everyone! I got it fixed and working for jonigl/mcp-client-for-ollama since v0.19.0.
However, I noticed that deepseek-v3.1:671b-cloud doesn’t seem to handle tool calls correctly. Tool calls aren’t being parsed as expected. I’ll share a screenshot to illustrate the issue. (I have opened I new issue just in case https://github.com/ollama/ollama/issues/12702)

Image
<!-- gh-comment-id:3421266789 --> @jonigl commented on GitHub (Oct 20, 2025): Hi everyone! I got it fixed and working for [jonigl/mcp-client-for-ollama](https://github.com/jonigl/mcp-client-for-ollama) since **v0.19.0**. However, I noticed that `deepseek-v3.1:671b-cloud` doesn’t seem to handle tool calls correctly. Tool calls aren’t being parsed as expected. I’ll share a screenshot to illustrate the issue. (I have opened I new issue just in case https://github.com/ollama/ollama/issues/12702) <img width="1421" height="185" alt="Image" src="https://github.com/user-attachments/assets/a9a4f0b7-b567-4e7d-9e5c-4300b31d0031" />
GiteaMirror commented 2026-04-22 17:12:48 -05:00
Author
Owner
Copy Link

@ParthSareen commented on GitHub (Nov 8, 2025):

Hi y'all. This error message pops up when there isn't a subsequent message with role tool after a message with tool_calls. The tool call the model made also needs to be sent to the model

<!-- gh-comment-id:3505541693 --> @ParthSareen commented on GitHub (Nov 8, 2025): Hi y'all. This error message pops up when there isn't a subsequent message with role `tool` after a message with `tool_calls`. The tool call the model made also needs to be sent to the model
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 cloud
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#33993
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?