BerriAI/litellm · error · ValueError

Missing required fields in the result={result}

Error message

Missing required fields in the result={result}

What it means

While iterating TogetherAI rerank results, LiteLLM requires each item in 'results' to contain at least 'index' and 'relevance_score'. If any element is missing either key (e.g. items reduced to {'document': ...} only, or null entries), this ValueError is raised with the offending result serialized in the message. It is a per-item schema validation of the provider payload.

Source

Thrown at litellm/llms/together_ai/rerank/transformation.py:36

class TogetherAIRerankConfig:
    def _transform_response(self, response: dict) -> RerankResponse:
        _billed_units: Final = RerankBilledUnits(**response.get("usage", {}))
        _tokens: Final = RerankTokens(**response.get("usage", {}))
        rerank_meta: Final = RerankResponseMeta(billed_units=_billed_units, tokens=_tokens)

        _results: Final[list[dict] | None] = response.get("results")

        if _results is None:
            raise ValueError(f"No results found in the response={response}")

        rerank_results: Final[list[RerankResponseResult]] = []

        for result in _results:
            # Validate required fields exist
            if not all(key in result for key in ["index", "relevance_score"]):
                raise ValueError(f"Missing required fields in the result={result}")

            # Get document data if it exists
            document_data = result.get("document", {})
            document = RerankResponseDocument(text=str(document_data.get("text", ""))) if document_data else None

            # Create typed result
            rerank_result = RerankResponseResult(
                index=int(result["index"]),
                relevance_score=float(result["relevance_score"]),
            )

            # Only add document if it exists
            if document:
                rerank_result["document"] = document

            rerank_results.append(rerank_result)

        return RerankResponse(

View on GitHub (pinned to 77b7c6c40c)

Solutions

1. Check the result object printed in the message to identify the missing field(s).
2. If you control the response source (mock/stub), include 'index' and 'relevance_score' in every result item.
3. Update litellm to the latest patch in case the provider's format changed and was adapted upstream.
4. If Together genuinely omits scores for your model, switch to a rerank model id documented to return them.

Example fix

# before — stub result missing required fields
# body: {"results": [{"document": {"text": "..."}}]}
resp = litellm.rerank(model="together_ai/rerank-english-v2.0", query=q, documents=docs)
# ValueError: Missing required fields in the result={'document': {'text': '...'}}

# after — each item carries index + relevance_score
# body: {"results": [
#   {"index": 0, "relevance_score": 0.97, "document": {"text": "..."}}
# ]}

def valid_rerank_results(results) -> bool:
    """Validate item shape before handing payloads to code that assumes it."""
    if not isinstance(results, list):
        return False
    return all(
        isinstance(r, dict) and "index" in r and "relevance_score" in r
        for r in results
    )

def is_complete_rerank_result(item: object) -> bool:
    """Type guard for a single Together rerank result item."""
    return (
        isinstance(item, dict)
        and isinstance(item.get("index"), int)
        and isinstance(item.get("relevance_score"), (int, float))
    )

try:
    resp = litellm.rerank(model="together_ai/rerank-english-v2.0", query=q, documents=docs)
except ValueError as e:
    if "Missing required fields in the result" in str(e):
        logger.error("together rerank item malformed: %s", e)
        resp = litellm.rerank(model="together_ai/rerank-english-v2.0", query=q, documents=docs[:50])
    else:
        raise

When it happens

Trigger: Together returns results entries lacking index or relevance_score (partial fields on certain models/modes); response-mocking tools that build plausible but incomplete items; downstream API version drift where scores are omitted (e.g. return_documents-only responses).

Common situations: Test fixtures hand-written from docs that skip optional-looking fields; Together A/B response formats; gateway transforms that strip fields; llm-judge pipelines consuming raw Together output.

Related errors

• No results found in the response={response}
• No results found in the response={response}
• Missing required fields in the result={result}
• TogetherAI does not support max_chunks_per_doc
• response.text

AI-assisted analysis of BerriAI/litellm@77b7c6c40c (2026-08-18). Data as JSON: /api/errors/9fd8c300ba6375d2. Report an issue: GitHub.