~ LOSSY

pydantic-ai 2.61.0

finish.unknown-value · ollama-openai

Requirements

Id	Level	Status	Evidence
response-returned	MUST	SATISFIED	outcome=returned; error.category=null; truth.error=null
finish-is-other	MUST	NOT_OBSERVABLE	finish.value=null (raw: provider_details.finish_reason); truth.finish="other"; required finish="other"
text-reported	MUST	SATISFIED	text.value="Octopuses, like many marine invertebrates, employ sophisticated methods to camouflage" (primary); truth.text="Octopuses, like many marine invertebrates, employ sophisticated methods to camouflage"
raw-finish-reported	SHOULD	SATISFIED	finish.raw="wiretruth_unknown_reason"; truth.raw_finish="wiretruth_unknown_reason"

What the provider sent

Finish: other
Raw finish: wiretruth_unknown_reason
Text: Octopuses, like many marine invertebrates, employ sophisticated methods to camouflage
Tool calls: []
Usage: { "alternatives": {}, "cached_input": 28, "input": 32, "output": 16, "raw": { "completion_tokens": 16, "prompt_tokens": 32, "prompt_tokens_details": { "cached_tokens": 28 }, "total_tokens": 48 }, "reasoning": null, "total": 48 }
Stream completed: null
Error: null

Wire excerpt

status: 200
content-type: application/json
body (json):
{
  "id": "chatcmpl-54",
  "object": "chat.completion",
  "created": 1791831683,
  "model": "qwen3.5:0.8b",
  "system_fingerprint": "fp_ollama",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Octopuses, like many marine invertebrates, employ sophisticated methods to camouflage"
      },
      "finish_reason": "wiretruth_unknown_reason"
    }
  ],
  "usage": {
    "prompt_tokens": 32,
    "prompt_tokens_details": {
      "cached_tokens": 28
    },
    "completion_tokens": 16,
    "total_tokens": 48
  }
}

What the library reported

Outcome

returned

Finish

{
  "raw": "wiretruth_unknown_reason",
  "source": "provider_details.finish_reason",
  "surface": "raw",
  "value": null
}

Text

Value: Octopuses, like many marine invertebrates, employ sophisticated methods to camouflage
Surface: primary

Tool calls

[]

Usage

{
  "cached_input": 28,
  "input": 32,
  "output": 16,
  "reasoning": 0,
  "surface": {
    "cached_input": "primary",
    "input": "primary",
    "output": "primary",
    "reasoning": "primary",
    "total": "primary"
  },
  "total": 48
}

Stream

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

Error

null

Unsupported reason

null

Full observation JSON

{
  "elapsed_ms": 170.20567040890455,
  "error": null,
  "finish": {
    "raw": "wiretruth_unknown_reason",
    "source": "provider_details.finish_reason",
    "surface": "raw",
    "value": null
  },
  "format": "wiretruth.observation/v1",
  "library": {
    "api": "Agent.run",
    "language": "python",
    "name": "pydantic-ai",
    "version": "2.61.0"
  },
  "notes": [],
  "outcome": "returned",
  "stream": {
    "chunks_seen": null,
    "completed": null,
    "used": false
  },
  "structured": {
    "attempted": false,
    "returned_as_success": false,
    "value": null
  },
  "text": {
    "surface": "primary",
    "value": "Octopuses, like many marine invertebrates, employ sophisticated methods to camouflage"
  },
  "tool_calls": [],
  "unsupported_reason": null,
  "usage": {
    "cached_input": 28,
    "input": 32,
    "output": 16,
    "reasoning": 0,
    "surface": {
      "cached_input": "primary",
      "input": "primary",
      "output": "primary",
      "reasoning": "primary",
      "total": "primary"
    },
    "total": 48
  }
}

Fixture

derived ollama-openai/finish.unknown-value

Provenance	Value
Transform	replace_finish_value
Source fixture	ollama-openai/truncation.text
Rationale	The provider sends a finish value that a library may not recognize.
Evidence	https://github.com/BerriAI/litellm/issues/21347

Reproduce

wiretruth serve ollama-openai/finish.unknown-value

Plain-library snippet from the shim manifest.

OllamaModel(model, provider=OllamaProvider(base_url=WIRETRUTH_BASE_URL_OLLAMA_OPENAI, api_key=WIRETRUTH_API_KEY)); /v1 base

Upstream

released Issue · PR

Fixed in: 2.61.0

Existing upstream issue, not opened by wiretruth; wiretruth's report, #10381, was closed as its duplicate. Up to 2.60.0, OpenAIChatModel re-validated the response with the strict ChatCompletion model (_validate_completion, models/openai.py), so a finish_reason outside OpenAI's five values raised UnexpectedModelBehavior and the text and usage were lost. OllamaModel inherits it. Since #11119 (2.61.0) the response is kept with finish_reason None and the raw value in provider_details['finish_reason'], so both cells move WRONG → LOSSY, the same as AnthropicModel, GoogleModel and OpenAIResponsesModel.

History

Version	Grade
2.61.0	~ LOSSY
2.56.0	✗ WRONG