"Python Dataclass vs NamedTuple: Which to Use and When"
"Choosing between a Python dataclass and a NamedTuple comes down to one question: does your record need to change or behave like a tuple? This comparison gives you the direct answer up front, a head-to-head table, and a verdict for API DTOs, config objects, and hot-path records."
Use a dataclass when you need mutability, defaults, validation, or methods; use a NamedTuple when you want a small, immutable, tuple-like record. That's the direct answer to the python dataclass vs namedtuple question, and it settles 80% of cases. The remaining 20% — frozen dataclasses that close the immutability gap, API boundaries where neither is the right tool, and one sneaky equality gotcha — is what this comparison is for. Below: the short answer with its exceptions, both options with runnable code, a full head-to-head table, and a verdict for every scenario from API DTOs to hot-path records. The short answer (and when it breaks down) The one-line rule works because the two types were designed for different jobs. NamedTuple is a tuple with names: fixed, immutable, unpackable. A dataclass is a class with generated boilerplate: flexible, mutable by default, extensible with methods and validation. The rule breaks down in two places worth knowing about: Frozen dataclasses close the gap. @dataclass(frozen=True, slots=True) gives you immutability, hashability, and no per-instance __dict__ — a near-NamedTuple with dataclass ergonomics. When someone says "dataclasses made NamedTuple obsolete," this is what they mean. (It's not the whole story — the table below shows what NamedTuple still wins.) API boundaries belong to neither. When you're parsing untrusted JSON into a validated model, reach for Pydantic at the edge and convert to a dataclass or NamedTuple internally — not for either of these two alone. If you want the deeper dataclass story first, /blog/python-oop-dataclasses covers it end to end. NamedTuple: the immutable tuple with names A NamedTuple is a subclass of tuple whose elements can be accessed by name. Prefer the typing.NamedTuple class syntax…