langchain-ai/langchain · error · ValueError

Got mismatched input_variables. Expected: {input_vars}. Got:

Error message

Got mismatched input_variables. Expected: {input_vars}. Got: {values['input_variables']}

What it means

ChatPromptTemplate's model validator computes input variables as the union across all message templates, minus partial and optional variables. If 'input_variables' was supplied explicitly AND values['validate_template'] is truthy, and the computed set differs from the supplied one, ValueError('Got mismatched input_variables...') is raised listing both sets. Without validate_template, the computed list silently overwrites the supplied one.

Source

Thrown at libs/core/langchain_core/prompts/chat.py:1097

                    message.optional
                    and message.variable_name not in values["partial_variables"]
                ):
                    values["partial_variables"][message.variable_name] = []
                    optional_variables.add(message.variable_name)
                if message.variable_name not in input_types:
                    input_types[message.variable_name] = list[AnyMessage]
        if "partial_variables" in values:
            input_vars -= set(values["partial_variables"])
        if optional_variables:
            input_vars -= optional_variables
        if "input_variables" in values and values.get("validate_template"):
            if input_vars != set(values["input_variables"]):
                msg = (
                    "Got mismatched input_variables. "
                    f"Expected: {input_vars}. "
                    f"Got: {values['input_variables']}"
                )
                raise ValueError(msg)
        else:
            values["input_variables"] = sorted(input_vars)
        if optional_variables:
            values["optional_variables"] = sorted(optional_variables)
        values["input_types"] = input_types
        return values

    @classmethod
    def from_template(cls, template: str, **kwargs: Any) -> ChatPromptTemplate:
        """Create a chat prompt template from a template string.

        Creates a chat template consisting of a single message assumed to be from the
        human.

        Args:
            template: Template string
            **kwargs: Keyword arguments to pass to the constructor.

View on GitHub (pinned to e32fa9a52e)

Solutions

Update the supplied input_variables to exactly match the variables used across message templates (partial/optional excluded)
Or omit input_variables entirely and let the validator derive them
Only set validate_template=True when you genuinely want strict cross-checking

Example fix

# before
ChatPromptTemplate(
    input_variables=["topic", "style"],
    messages=["human: Tell me about {topic}"],
    validate_template=True,
)  # ValueError: expected {'topic'}

# after
ChatPromptTemplate(
    input_variables=["topic"],
    messages=["human: Tell me about {topic}"],
    validate_template=True,
)

Defensive patterns

Strategy: validation

Validate before calling

def derive_input_vars(messages) -> set[str]:
    vars_ = set()
    for m in messages:
        vars_ |= set(getattr(m, "input_variables", []))
    return vars_

# before constructing with validate_template=True:
assert set(declared) == derive_input_vars(messages)

Prevention

Omit input_variables and let the validator derive them
When editing template strings, update (or drop) explicit input_variables lists in the same change

When it happens

Trigger: Constructing ChatPromptTemplate(input_variables=['a','b'], messages=[...]) with validate_template=True where the templates actually reference {'a','c'}; typical when hand-maintained input_variables lists drift from template edits.

Common situations: Copy-pasted constructors where the template string changed but the explicit input_variables list did not; migration from legacy ChatPromptTemplate(prompt=...) kwargs that required manual variable lists.

Related errors

AI-assisted analysis of langchain-ai/langchain@e32fa9a52e (2026-08-14). Data as JSON: /api/errors/335a95cc74318206. Report an issue: GitHub.