BerriAI/litellm · error · ValueError

No results found in the response={response}

Error message

No results found in the response={response}

What it means

After a successful TogetherAI rerank HTTP call, LiteLLM parses the JSON body and expects a 'results' key. If response.get('results') is None — the key is absent or explicitly null — _transform_response raises ValueError with the whole response embedded, because it cannot construct a RerankResponse without results. This indicates a malformed or unexpected Together payload rather than an HTTP failure (that would have raised earlier).

Source

Thrown at litellm/llms/together_ai/rerank/transformation.py:29

    RerankBilledUnits,
    RerankResponse,
    RerankResponseDocument,
    RerankResponseMeta,
    RerankResponseResult,
    RerankTokens,
)


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"]),
            )

View on GitHub (pinned to 77b7c6c40c)

Solutions

1. Upgrade (or pin) litellm to a release matching the current Together rerank response schema.
2. Inspect the response echoed in the message to see what Together actually returned (often an inline error object).
3. If a mock/proxy intercepts the call, make it return {'results': [...], 'usage': {...}} shaped payloads.
4. Report the exact response body to litellm maintainers if the provider changed its contract.

Example fix

# before (mock/interceptor returns wrong shape)
# 200 OK body: {"message": "ok"}
resp = litellm.rerank(model="together_ai/rerank-english-v2.0", query=q, documents=docs)
# ValueError: No results found in the response={'message': 'ok'}

# after — return the documented shape
# 200 OK body: {"results": [{"index": 0, "relevance_score": 0.98}], "usage": {...}}
resp = litellm.rerank(model="together_ai/rerank-english-v2.0", query=q, documents=docs)

def looks_like_together_rerank_response(body: dict) -> bool:
    """Cheap contract check when you control/mock the response source."""
    return isinstance(body, dict) and isinstance(body.get("results"), list)

try:
    resp = litellm.rerank(model="together_ai/rerank-english-v2.0", query=q, documents=docs)
except ValueError as e:
    if "No results found in the response" in str(e):
        # provider contract break: log body for maintainers, fall back to no rerank
        logger.warning("together rerank returned no results: %s", e)
        resp = None
    else:
        raise

When it happens

Trigger: Together returns 200 with a body lacking 'results' (e.g. an error object, empty response, or an API shape change); a proxy/interceptor rewrites the response body; a different endpoint version returns {'message': ...} with status 200.

Common situations: Together shipping a response-format change that LiteLLM's current version doesn't handle; middlewares (mock servers, gateway rewrites) returning stub JSON; edge cases where Together returns an empty result set as null.

Related errors

• Missing required fields in the result={result}
• 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/eead51e2c267cbe8. Report an issue: GitHub.