Implementation:Bentoml BentoML Models List Get

Implementation Metadata
Implementation Name	Models List Get
API	bentoml.models.list(), bentoml.models.get()
Source	src/bentoml/models.py:L31-49
Workflow	Model_Store_Management
Domain	ML_Serving, Model_Management
Implements	Principle:Bentoml_BentoML_Model_Versioning
Last Updated	2026-02-13 15:00 GMT

Overview

The bentoml.models.list() and bentoml.models.get() functions provide querying and retrieval capabilities for models in the BentoML local store. list() returns all models (optionally filtered by name), while get() retrieves a single model by its exact tag.

Import

import bentoml

Signatures

def list(tag: Tag | str = None) -> list[Model]

def get(tag: Tag | str) -> Model

Parameters

bentoml.models.list()

Parameter	Type	Default	Description
tag	str	None	Optional filter. If a name is provided (e.g., "my_model"), only versions of that model are returned. If None, all models in the store are returned.

bentoml.models.get()

Parameter	Type	Default	Description
tag	str	required	The model tag to retrieve. Supports "name" (resolves to latest), "name:latest", or "name:version".

Inputs and Outputs

list()

Inputs:

• Optional tag filter (model name or full tag)

Outputs:

• list[Model] sorted by creation time (newest first)
• Each Model object contains:
  • .tag — the immutable name:version tag
  • .info — structured model info including labels, metadata, creation_time
  • .path — filesystem path to the model artifacts

get()

Inputs:

• Exact tag (or name for latest resolution)

Outputs:

• Single Model instance
• Raises NotFound if the model does not exist in the store

Usage Examples

import bentoml

# List all models in the store
all_models = bentoml.models.list()
for model in all_models:
    print(f"{model.tag} - created: {model.info.creation_time}")

# List all versions of a specific model
classifier_versions = bentoml.models.list("text_classifier")

# Get a specific model version
model = bentoml.models.get("text_classifier:abc123")
print(f"Path: {model.path}")
print(f"Labels: {model.info.labels}")
print(f"Metadata: {model.info.metadata}")

# Get the latest version of a model
latest = bentoml.models.get("text_classifier:latest")
# Or equivalently:
latest = bentoml.models.get("text_classifier")

Model Info Structure

The Model.info object returned by both functions contains:

Field	Type	Description
tag	Tag	The immutable model tag (name:version)
labels	dict[str, str]	Key-value labels for filtering
metadata	dict[str, Any]	Arbitrary user metadata
creation_time	datetime	When the model was saved
context	ModelContext	Framework and Python version info
module	str	The module that saved the model

Behavior Details

• list() returns results sorted by creation time with the newest models first.
• get() supports the "latest" version specifier, which resolves to the most recently saved version of a given model name.
• Both functions operate on the local store only. They do not query BentoCloud; use BentoModel.resolve() for automatic cloud fallback.

Source Reference

File: src/bentoml/models.py, lines 31-49.

Knowledge Sources

• BentoML
• BentoML Docs