github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-22 21:51:37 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #15647] tool_call.arguments truncated to exactly 214 chars on Ollama Cloud API #72043

New Issue
Open
opened 2026-05-05 03:23:40 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-05-05 03:23:40 -05:00
Owner
Copy Link

Originally created by @ruhtraxjj on GitHub (Apr 17, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/15647

Bug Description
When using models via the Ollama Cloud API (https://ollama.com/v1), tool_call.function.arguments is truncated to exactly 214 characters.
This produces invalid JSON that causes HTTP 400: invalid tool call
arguments on subsequent API calls that include the truncated
arguments in conversation history.

Steps to Reproduce

  1. Call POST /v1/chat/completions with glm-5.1 (or qwen3.5) and tool definitions
  2. In a multi-turn conversation where the model generates tool calls
    with arguments > 214 chars 3. The returned tool_call.function.arguments is truncated to exactly
    214 characters 4. Sending the conversation history (including the truncated args)
    back to the API returns 400: invalid tool call arguments
    Evidence
    Across 35+ error instances, every single invalid arguments field is exactly 214 characters long, strongly suggesting a server-side buffer or limit:

All 46 invalid tool_calls: 214 chars (100%)

Examples of truncated arguments:

  • {"code":"from hermes_tools import terminal\n\nrepos = [\n
    ("alexpate/awesome-design-systems"... → truncated mid-string - {"command":"cd ~/design-md && for d in */; do echo "${d%/}"; done
    | sort > /tmp/... → truncated mid-command - {"path":"/home/emma/北雅英语/BEIYA-DESIGN-WORKFLOW.md","content":"# 北雅 AI 设计... → truncated mid-content
    Expected Behavior
    tool_call.function.arguments should contain complete, valid JSON regardless of length. If there is a length limit, the model should be constrained to produce arguments within that limit, or the API
    should return an error rather than silently truncating.

Environment

  • Ollama Cloud API (https://ollama.com/v1) - Models: glm-5.1, qwen3.5:397b, glm-5.1:cloud (via local Ollama)
  • Plan: Pro
Originally created by @ruhtraxjj on GitHub (Apr 17, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/15647 Bug Description When using models via the Ollama Cloud API (https://ollama.com/v1), tool_call.function.arguments is truncated to exactly 214 characters. This produces invalid JSON that causes HTTP 400: invalid tool call arguments on subsequent API calls that include the truncated arguments in conversation history. Steps to Reproduce 1. Call POST /v1/chat/completions with glm-5.1 (or qwen3.5) and tool definitions 2. In a multi-turn conversation where the model generates tool calls with arguments > 214 chars 3. The returned tool_call.function.arguments is truncated to exactly 214 characters 4. Sending the conversation history (including the truncated args) back to the API returns 400: invalid tool call arguments Evidence Across 35+ error instances, every single invalid arguments field is exactly 214 characters long, strongly suggesting a server-side buffer or limit: All 46 invalid tool_calls: 214 chars (100%) Examples of truncated arguments: - {"code":"from hermes_tools import terminal\n\nrepos = [\n (\"alexpate/awesome-design-systems\"... → truncated mid-string - {"command":"cd ~/design-md && for d in */; do echo \"${d%/}\"; done | sort > /tmp/... → truncated mid-command - {"path":"/home/emma/北雅英语/BEIYA-DESIGN-WORKFLOW.md","content":"# 北雅 AI 设计... → truncated mid-content Expected Behavior tool_call.function.arguments should contain complete, valid JSON regardless of length. If there is a length limit, the model should be constrained to produce arguments within that limit, or the API should return an error rather than silently truncating. Environment - Ollama Cloud API (https://ollama.com/v1) - Models: glm-5.1, qwen3.5:397b, glm-5.1:cloud (via local Ollama) - Plan: Pro
GiteaMirror added the cloud label 2026-05-05 03:23:40 -05:00
GiteaMirror commented 2026-05-05 03:23:41 -05:00
Author
Owner
Copy Link

@PureBlissAK commented on GitHub (Apr 18, 2026):

🤖 Automated Triage & Analysis Report

Issue: #15647
Analyzed: 2026-04-18T18:13:50.898637

Analysis

  • Type: unknown
  • Severity: medium
  • Components: unknown

Implementation Plan

  • Effort: medium
  • Steps:

This issue has been triaged and marked for implementation.

<!-- gh-comment-id:4274295028 --> @PureBlissAK commented on GitHub (Apr 18, 2026): <!-- ollama-issue-orchestrator:v1 issue:15647 --> ## 🤖 Automated Triage & Analysis Report **Issue**: #15647 **Analyzed**: 2026-04-18T18:13:50.898637 ### Analysis - **Type**: unknown - **Severity**: medium - **Components**: unknown ### Implementation Plan - **Effort**: medium - **Steps**: *This issue has been triaged and marked for implementation.*
GiteaMirror commented 2026-05-05 03:23:42 -05:00
Author
Owner
Copy Link

@horendus commented on GitHub (Apr 21, 2026):

Solid analysis by the bot

<!-- gh-comment-id:4289509708 --> @horendus commented on GitHub (Apr 21, 2026): Solid analysis by the bot
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 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#72043
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?