Model retirements · checked

Model deprecations and shutdown dates

On 8 October 2026, the official deprecation pages of OpenAI, Anthropic and Google listed 27 language models due to shut down by 6 January 2027; the next is gemini-2.5-flash on Vertex AI (Gemini Enterprise Agent Platform), on 20 October 2026. Each one links to its replacement, the price ratio at list prices and the API changes the provider documents. Record a model's answers on your own prompts before its date: after it, nobody can.

$ npx --allow-git=root github:agentwares/model-migrate scan

npx --allow-git=root github:agentwares/model-migrate scan lists every model ID in your code and config that has a shutdown date, with its replacement, price ratio and this page's link. It needs no account and no API key, reads your files on your machine and sends nothing; its one request fetches this data. Free and MIT licensed.

Shutting down

ShutdownModelPlatformReplacementPrice per token, in / out
20 Oct 2026gemini-2.5-flashVertex AI (Gemini Enterprise Agent Platform)gemini-3.8-flash or gemini-3.5-flash-lite or gemini-3.1-flash-lite2.5× / 1.5× until 31 Dec 2026
20 Oct 2026gemini-2.5-flash-liteVertex AI (Gemini Enterprise Agent Platform)gemini-3.8-flash or gemini-3.1-flash-lite or Gemma 47.5× / 9.4× until 31 Dec 2026
20 Oct 2026gemini-2.5-proVertex AI (Gemini Enterprise Agent Platform)gemini-3.8-flash or gemini-3.5-flash0.6× / 0.38× until 31 Dec 2026
23 Oct 2026ft-babbage-002OpenAI APIgpt-5.6-terranot listed
23 Oct 2026ft-davinci-002OpenAI APIgpt-5.6-terranot listed
23 Oct 2026ft-gpt-3.5-turboOpenAI APIgpt-5.6-terranot listed
23 Oct 2026ft-gpt-4OpenAI APIgpt-5.6-solnot listed
23 Oct 2026ft-gpt-4.1-nano-2025-04-14OpenAI APIgpt-5.6-lunanot listed
23 Oct 2026ft-o4-mini-2025-04-16OpenAI APIgpt-5.6-terranot listed
23 Oct 2026gpt-3.5-turboOpenAI APIgpt-5.6-terra4× / 8×
23 Oct 2026gpt-4OpenAI APIgpt-5.6-sol0.13× / 0.33×
23 Oct 2026gpt-4-1106-previewOpenAI APIgpt-5.6-solnot listed
23 Oct 2026gpt-4-turboOpenAI APIgpt-5.6-sol0.4× / 0.67×
23 Oct 2026gpt-4.1-nanoOpenAI APIgpt-5.6-luna2× / 3×
23 Oct 2026gpt-4o-2024-05-13OpenAI APIgpt-5.6-sol0.8× / 1.3×
23 Oct 2026o1OpenAI APIgpt-5.6-sol0.27× / 0.33×
23 Oct 2026o1-proOpenAI APIgpt-5.6-solnot listed
23 Oct 2026o3-miniOpenAI APIgpt-5.6-sol3.6× / 4.6×
23 Oct 2026o4-miniOpenAI APIgpt-5.6-terra1.8× / 2.7×
19 Nov 2026gemini-3.6-flashVertex AI (Gemini Enterprise Agent Platform)gemini-3.8-flash1× / 1×
30 Nov 2026claude-sonnet-4-5Claude APIclaude-sonnet-5-50.67× / 0.67×
11 Dec 2026gpt-5OpenAI APIgpt-5.6-sol3.2× / 2×
11 Dec 2026gpt-5-miniOpenAI APIgpt-5.6-terra8× / 6×
11 Dec 2026gpt-5-nanoOpenAI APIgpt-5.6-luna4× / 3×
11 Dec 2026gpt-5-proOpenAI APIgpt-5.6-solnot listed
11 Dec 2026o3OpenAI APIgpt-5.6-sol2× / 2.5×
11 Dec 2026o3-proOpenAI APIgpt-5.6-solnot listed
28 Jan 2027gemini-3.7-flashVertex AI (Gemini Enterprise Agent Platform)gemini-3.8-flash1× / 1×
1 Apr 2027gpt-5.1OpenAI APIgpt-6-sol1.6× / 1×
1 Apr 2027gpt-5.3-codexOpenAI APIgpt-6-sol1.1× / 0.71×
1 Apr 2027gpt-5.4-nanoOpenAI APIgpt-6-luna0.5× / 0.4×
from 7 May 2027gemini-3.1-flash-liteGemini APIgemini-3.5-flash-lite1.2× / 1.7×

Already shut down

Requests to these models now fail. The pages stay up for anyone migrating late.

ShutdownModelPlatformReplacementPrice per token, in / out
1 Oct 2026gpt-5.4-cyberOpenAI APInone namednot listed
28 Sep 2026babbage-002OpenAI APIgpt-5.6-terra5× / 30×
28 Sep 2026davinci-002OpenAI APIgpt-5.6-terra1× / 6×
28 Sep 2026gpt-3.5-turbo-1106OpenAI APIgpt-5.6-terra2× / 6×
28 Sep 2026gpt-3.5-turbo-instructOpenAI APIgpt-5.6-terra1.3× / 6×
10 Aug 2026gpt-5.2-chat-latestOpenAI APIgpt-5.6-solnot listed
10 Aug 2026gpt-5.3-chat-latestOpenAI APIgpt-5.6-solnot listed
5 Aug 2026claude-opus-4-1Claude APIclaude-opus-4-80.33× / 0.33×
from 28 Jul 2026gemini-2.5-computer-use-preview-10-2025Gemini APIgemini-3.8-flashnot listed
23 Jul 2026computer-use-previewOpenAI APIgpt-5.6-terranot listed
23 Jul 2026gpt-4o-mini-search-preview-2025-03-11OpenAI APIgpt-5.6-terranot listed
23 Jul 2026gpt-4o-search-preview-2025-03-11OpenAI APIgpt-5.6-terranot listed
23 Jul 2026gpt-5-chat-latestOpenAI APIgpt-5.6-solnot listed
23 Jul 2026gpt-5-codexOpenAI APIgpt-5.6-solnot listed
23 Jul 2026gpt-5.1-chat-latestOpenAI APIgpt-5.6-solnot listed
23 Jul 2026gpt-5.1-codexOpenAI APIgpt-5.6-solnot listed
23 Jul 2026gpt-5.1-codex-maxOpenAI APIgpt-5.6-solnot listed
23 Jul 2026gpt-5.1-codex-miniOpenAI APIgpt-5.6-terranot listed
23 Jul 2026gpt-5.2-codexOpenAI APIgpt-5.6-solnot listed
23 Jul 2026o3-deep-researchOpenAI APIgpt-5.6-solnot listed
23 Jul 2026o4-mini-deep-researchOpenAI APIgpt-5.6-solnot listed
15 Jun 2026claude-opus-4-20250514Claude APIclaude-opus-4-80.33× / 0.33×
15 Jun 2026claude-sonnet-4-20250514Claude APIclaude-sonnet-4-61× / 1×
from 1 Jun 2026gemini-2.0-flashGemini APIgemini-3.6-flashnot listed
from 1 Jun 2026gemini-2.0-flash-001Gemini APIgemini-3.6-flashnot listed
from 1 Jun 2026gemini-2.0-flash-liteGemini APIgemini-3.1-flash-litenot listed
from 1 Jun 2026gemini-2.0-flash-lite-001Gemini APIgemini-3.1-flash-litenot listed
1 Jun 2026gemini-2.0-flashVertex AI (Gemini Enterprise Agent Platform)gemini-3.1-flash-litenot listed
1 Jun 2026gemini-2.0-flash-liteVertex AI (Gemini Enterprise Agent Platform)gemini-3.1-flash-litenot listed
from 25 May 2026gemini-3.1-flash-lite-previewGemini APIgemini-3.1-flash-litenot listed
20 Apr 2026claude-3-haiku-20240307Claude APIclaude-haiku-4-5-20251001not listed
from 31 Mar 2026gemini-2.5-flash-lite-preview-09-2025Gemini APIgemini-3.1-flash-litenot listed
26 Mar 2026gpt-4-0125-previewOpenAI APIgpt-5 or gpt-4.1not listed
26 Mar 2026gpt-4-0314OpenAI APIgpt-5 or gpt-4.1not listed
from 9 Mar 2026gemini-3-pro-previewGemini APIgemini-3.1-pro-previewnot listed
19 Feb 2026claude-3-5-haiku-20241022Claude APIclaude-haiku-4-5-202510011.3× / 1.3×
19 Feb 2026claude-3-7-sonnet-20250219Claude APIclaude-sonnet-4-6not listed
from 17 Feb 2026gemini-2.5-flash-preview-09-25Gemini APIgemini-3.6-flashnot listed
17 Feb 2026chatgpt-4o-latestOpenAI APIgpt-5.1-chat-latestnot listed
12 Feb 2026codex-mini-latestOpenAI APIgpt-5-codex-mininot listed
5 Jan 2026claude-3-opus-20240229Claude APIclaude-opus-4-8not listed
from 9 Dec 2025gemini-2.0-flash-lite-previewGemini APIgemini-2.5-flash-litenot listed
from 9 Dec 2025gemini-2.0-flash-lite-preview-02-05Gemini APIgemini-2.5-flash-litenot listed
from 2 Dec 2025gemini-2.5-pro-preview-03-25Gemini APIgemini-3.1-pro-previewnot listed
from 2 Dec 2025gemini-2.5-pro-preview-05-06Gemini APIgemini-3.1-pro-previewnot listed
from 2 Dec 2025gemini-2.5-pro-preview-06-05Gemini APIgemini-3.1-pro-previewnot listed
from 18 Nov 2025gemini-2.5-flash-preview-05-20Gemini APIgemini-3.6-flashnot listed
28 Oct 2025claude-3-5-sonnet-20240620Claude APIclaude-sonnet-4-6not listed
28 Oct 2025claude-3-5-sonnet-20241022Claude APIclaude-sonnet-4-6not listed
27 Oct 2025o1-miniOpenAI APIo4-mininot listed
24 Sep 2025gemini-1.5-flash-002Vertex AI (Gemini Enterprise Agent Platform)gemini-2.5-flash-litenot listed
24 Sep 2025gemini-1.5-pro-002Vertex AI (Gemini Enterprise Agent Platform)gemini-2.5-flashnot listed
28 Jul 2025o1-previewOpenAI APIo3not listed
21 Jul 2025claude-2.0Claude APIclaude-opus-4-8not listed
21 Jul 2025claude-2.1Claude APIclaude-opus-4-8not listed
21 Jul 2025claude-3-sonnet-20240229Claude APIclaude-sonnet-4-6not listed
14 Jul 2025gpt-4.5-previewOpenAI APIgpt-4.1not listed

What is covered

Language models (chat, reasoning, completions, code, search, computer-use and deep-research IDs, and OpenAI fine-tune families) with a shutdown date on or after 1 July 2025, from four official pages: OpenAI's deprecations page, Anthropic's model deprecations doc, the Gemini API deprecations page and Vertex AI's model versions page. Audio, realtime, speech, transcription, image, video, embedding and moderation models are left out.

US dollars per million tokens at the standard tier, text input, shortest context tier, as listed on the provider's pricing page on the date seen. A ratio is the replacement's price divided by the retiring model's; it ignores differences in how many tokens each model uses.

Built only from these pages; no third-party calendar. The whole list is at /models/retirements.json and from the MCP tool agentcheck_model_retirements on agentcheck's server.

model-migrate is a free tool from agentcheck, which re-runs an agent's checks when a provider ships a model and replays recorded traces nightly, so a model change that breaks it arrives as one diff by email.