✓ PASS

litellm 1.105.5

truncation.reasoning-only · ollama

Requirements

Id	Level	Status	Evidence
cause-is-truncation	MUST	SATISFIED	finish.value="length" (primary: choices[0].finish_reason); truth.finish="length"; error.category=null; structured.returned_as_success=false
text-reported	MUST	SATISFIED	text.value="" (primary); truth.text=""
usage-reported	SHOULD	SATISFIED	usage.output=64 (primary); truth.usage.output=64

What the provider sent

Finish: length
Raw finish: length
Text
Tool calls: []
Usage: { "alternatives": {}, "cached_input": 0, "input": 54, "output": 64, "raw": { "eval_count": 64, "prompt_eval_cached_count": 0, "prompt_eval_count": 54 }, "reasoning": null, "total": 118 }
Stream completed: null
Error: null

Wire excerpt

status: 200
content-type: application/json; charset=utf-8
body (json):
{
  "model": "qwen3.5:0.8b",
  "created_at": "2026-10-12T19:01:17.433604074Z",
  "message": {
    "role": "assistant",
    "content": "",
    "thinking": "Let $n$ be the smallest positive whole number satisfying the given conditions.\nThe conditions are:\n1. Remainder when dividing by 2 is 1. So $n \\equiv 1 \\pmod 2$. This implies $n$ is odd.\n2. Remainder when dividing by"
  },
  "done": true,
  "done_reason": "length",
  "total_duration": 3728486452,
  "load_duration": 2447425,
  "prompt_eval_count": 54,
  "prompt_eval_cached_count": 0,
  "prompt_eval_duration": 554808000,
  "eval_count": 64,
  "eval_duration": 3162410000
}

What the library reported

Outcome

returned

Finish

{
  "raw": "length",
  "source": "choices[0].finish_reason",
  "surface": "primary",
  "value": "length"
}

Text

Value
Surface: primary

Tool calls

[]

Usage

{
  "cached_input": null,
  "input": 54,
  "output": 64,
  "reasoning": null,
  "surface": {
    "cached_input": "absent",
    "input": "primary",
    "output": "primary",
    "reasoning": "absent",
    "total": "primary"
  },
  "total": 118
}

Stream

{
  "chunks_seen": null,
  "completed": null,
  "used": false
}

Error

null

Unsupported reason

null

Full observation JSON

{
  "elapsed_ms": 334.1834330931306,
  "error": null,
  "finish": {
    "raw": "length",
    "source": "choices[0].finish_reason",
    "surface": "primary",
    "value": "length"
  },
  "format": "wiretruth.observation/v1",
  "library": {
    "api": "litellm.completion",
    "language": "python",
    "name": "litellm",
    "version": "1.105.5"
  },
  "notes": [],
  "outcome": "returned",
  "stream": {
    "chunks_seen": null,
    "completed": null,
    "used": false
  },
  "structured": {
    "attempted": false,
    "returned_as_success": false,
    "value": null
  },
  "text": {
    "surface": "primary",
    "value": ""
  },
  "tool_calls": [],
  "unsupported_reason": null,
  "usage": {
    "cached_input": null,
    "input": 54,
    "output": 64,
    "reasoning": null,
    "surface": {
      "cached_input": "absent",
      "input": "primary",
      "output": "primary",
      "reasoning": "absent",
      "total": "primary"
    },
    "total": 118
  }
}

Fixture

recorded ollama/truncation.reasoning-only

Provenance	Value
Recorded at	2026-10-12T19:01:13Z
Requested model	qwen3.5:0.8b
Returned model	qwen3.5:0.8b
CI run	https://github.com/kartsan03/wiretruth/actions/runs/38294709109

Reproduce

wiretruth serve ollama/truncation.reasoning-only

Plain-library snippet from the shim manifest.

litellm.completion(model='ollama_chat/' + model, api_base=WIRETRUTH_BASE_URL_OLLAMA, api_key=WIRETRUTH_API_KEY); host root

Upstream

No matching upstream report.

History

Version	Grade
1.105.5	✓ PASS
1.105.1	✓ PASS