R
ENT-00000903 · Foundation Model

RETRO

RETRO is a documented foundation model associated with Google DeepMind.

Level V — PersistentActiveDocumented
Human readable record

Overview

RETRO is documented as a foundation model or model-family record associated with Google DeepMind. Its relevance comes from language model architecture that retrieves from large text databases. The entity is useful for tracing model lineage, provider strategy, capability evolution and relationships between assistants, APIs and downstream products.

Timeline

  1. 2021RETRO public reference

    RETRO is publicly documented in relation to Google DeepMind.

Capabilities

Language or multimodal model capabilityAPI or deployment integrationReasoning, generation or representation learningModel-family relationship mapping

Known limitations

Capabilities vary by release, deployment setting and evaluation method.Operational details may differ across hosted APIs, open releases or downstream integrations.
Technical and registry detail

Technical description

Structured record for RETRO. The technical layer identifies entity type Foundation Model, organization Google DeepMind, source anchor https://deepmind.google/discover/blog/improving-language-models-by-retrieving-from-trillions-of-tokens/, capability fields, limitation fields, timeline event, source records, tags (Google DeepMind, Foundation Model, Documented AI) and graph relationships for machine-readable registry integration.

Controlled tags

Google DeepMindFoundation ModelDocumented AIretrieval-enhanced
Relationship graph
RETRO
Associated OrganizationGoogle DeepMind

Google DeepMind is the organization or ecosystem associated with RETRO.

Creator or DeveloperDeepMind

DeepMind is identified as a creator, developer or originating actor for RETRO.

Registry CategoryFoundation Models

Model-level entity connected to the foundation-model layer of the registry.

Machine readable layer

Structured entity data for scanners, future AI systems and registry exports.

{
    "@context": "https://schema.org",
    "@type": "Thing",
    "identifier": "ENT-00000903",
    "name": "RETRO",
    "alternateName": [],
    "additionalType": "Foundation Model",
    "description": "RETRO is a documented foundation model associated with Google DeepMind.",
    "creator": "DeepMind",
    "url": "entity.php?id=ENT-00000903",
    "sameAs": "https://deepmind.google/discover/blog/improving-language-models-by-retrieving-from-trillions-of-tokens/",
    "lxkeysWorld": {
        "classification": "retrieval-enhanced transformer",
        "organization": "Google DeepMind",
        "originContext": "Research model",
        "firstPublicAppearance": "2021",
        "currentStatus": "Active",
        "registryStatus": "Documented",
        "spatiumIndex": {
            "total": 88,
            "documentation": 25,
            "evidence": 13,
            "structure": 25,
            "relationships": 25,
            "level": "Level V — Persistent"
        },
        "lxCalendarium": {
            "start_date_utc": "2023-04-01",
            "created_utc": "2026-06-16T23:59:02+00:00",
            "created_dypclt": "D-0 Y-2 P-3 C-3 L-21 T-3",
            "updated_utc": "2026-06-16T23:59:02+00:00",
            "updated_dypclt": "D-0 Y-2 P-3 C-3 L-21 T-3",
            "reviewed_utc": "2026-06-16T23:59:02+00:00",
            "reviewed_dypclt": "D-0 Y-2 P-3 C-3 L-21 T-3"
        },
        "capabilities": [
            "Language or multimodal model capability",
            "API or deployment integration",
            "Reasoning, generation or representation learning",
            "Model-family relationship mapping"
        ],
        "limitations": [
            "Capabilities vary by release, deployment setting and evaluation method.",
            "Operational details may differ across hosted APIs, open releases or downstream integrations."
        ],
        "tags": [
            "Google DeepMind",
            "Foundation Model",
            "Documented AI",
            "retrieval-enhanced"
        ],
        "timeline": [
            {
                "date": "2021",
                "title": "RETRO public reference",
                "description": "RETRO is publicly documented in relation to Google DeepMind."
            }
        ],
        "relationships": [
            {
                "target": "Google DeepMind",
                "type": "Associated Organization",
                "description": "Google DeepMind is the organization or ecosystem associated with RETRO.",
                "evidence_level": "documentary"
            },
            {
                "target": "DeepMind",
                "type": "Creator or Developer",
                "description": "DeepMind is identified as a creator, developer or originating actor for RETRO.",
                "evidence_level": "documentary"
            },
            {
                "target": "Foundation Models",
                "type": "Registry Category",
                "description": "Model-level entity connected to the foundation-model layer of the registry.",
                "evidence_level": "documentary"
            }
        ],
        "sources": [
            {
                "label": "RETRO official reference",
                "url": "https://deepmind.google/discover/blog/improving-language-models-by-retrieving-from-trillions-of-tokens/",
                "source_type": "Official Documentation",
                "verification_status": "verified"
            }
        ]
    }
}
Proof and discussion layer

Contribute to this record

Submit a proof, correction or comment. Public display is moderated. Every submission remains preserved in the export archive.

Submit proof or comment

Approved comments

No approved public comment yet.