AI Model Lifecycle & Deprecation Index

A source-cited, cross-vendor record of AI model lifecycle — is a model still live, when does it sunset, and what's the migration target — across OpenAI, Anthropic, Google (Gemini), and Amazon Bedrock. Every date is transcribed from the provider's own official deprecation/lifecycle page and carries that source URL. Built for agents that need to know, at runtime, whether the model they're about to call still exists.

• Site: https://data.oscar-lopez.com/model-lifecycle
• API + MCP host: https://lifecycle.oscar-lopez.com
• Maintainer: Oscar E. López, Ph.D. (Information Systems, UMass Amherst) — independent project.
• License: CC BY 4.0 · Cite as: López, O. E. (2026). AI Model Lifecycle & Deprecation Index. data.oscar-lopez.com/model-lifecycle

Coverage

71 models across 4 providers (active / preview / deprecated / retired), with status, release / deprecation / shutdown dates, and provider-recommended replacements. Verified 2026-09-08.

REST endpoints (free)

• GET /v1/lifecycle/model?id=ID — one model: status, dates, days-to-shutdown, replacement, source
• GET /v1/lifecycle/deprecations?within_days=90 — models shutting down within N days
• GET /v1/lifecycle/provider?provider=openai|anthropic|google|aws-bedrock — a provider's models
• GET /v1/lifecycle/stats — counts by provider/status + next shutdown
• GET /v1/lifecycle/all — full export (free during the public preview; metered via x402 when live)

MCP tools

Endpoint: https://lifecycle.oscar-lopez.com/mcp

• model_lifecycle — full lifecycle of one model (free)
• is_model_live — is a model still callable, with its sunset date (free)
• upcoming_deprecations — models sunsetting within N days across all providers (free)
• get_model_lifecycle_all — full export (free during preview)

Discovery

/.well-known/agent-card.json (A2A), /.well-known/x402, /openapi.json

How it's built

Cloudflare Worker + D1. The index is derived live from a models table; days-until-shutdown is computed at request time. Values come from official provider pages only; fields a provider doesn't publish are left blank rather than inferred. A model's lifecycle on a hosting platform (e.g. Amazon Bedrock) can differ from the model-maker's own API — each row records which surface it describes.