LLM Status

AI model deprecation tracker

Every major provider's model deprecations and retirements — with shutdown dates and the recommended replacement for each — in one place, updated continuously. Currently tracking 66 upcoming retirements across 7 providers.

See if your code uses any of these — one scan, no account:

curl -fsSL https://llmstatus.ai/install.sh | bash

Upcoming retirements · 66

Llama 4 Maverick 17B 128E Instruct FP8
Microsoft · llama-4-maverick-17b-128e-instruct-fp8
2026-11-19
today
GLM-5
Z.ai (Zhipu)
2026-11-20
in 1d
Claude Opus 4.5
Anthropic · claude-opus-4-5-20251101
2026-11-24
in 5d
Gemini 2.5 Flash → Gemini 3.5 Flash Lite
Google · gemini-2.5-flash
2026-12-01
in 12d
Kimi K2.6
Moonshot AI · kimi-k2.6
2026-12-01
in 12d
Kimi K2.7 Code HighSpeed
Moonshot AI · kimi-k2.7-code-highspeed
2026-12-01
in 12d
Kimi K2.7 Code
Moonshot AI · kimi-k2.7-code
2026-12-01
in 12d
chatgpt-image-latest → GPT Image 2.5 Sunburst
OpenAI
2026-12-01
in 12d
gpt-image-1.5 → GPT Image 2.5 Sunburst
OpenAI
2026-12-01
in 12d
Gemini 2.5 Pro → Gemini 3.8 Flash
Google · gemini-2.5-pro
2026-12-01
in 12d
gpt-image-1-mini → GPT Image 2.5 Sunburst
OpenAI
2026-12-01
in 12d
Ministral 3B
Microsoft · ministral-3b
2026-12-09
in 20d
Mistral Medium 3
Microsoft · mistral-medium-2505
2026-12-09
in 20d
o3
Microsoft
2026-12-11
in 22d
GPT-5 Nano
Microsoft · gpt-5-nano
2026-12-11
in 22d
OpenAI o3 → GPT-5.5
OpenAI · o3-2025-04-16
2026-12-11
in 22d
GPT-5 → GPT-5.5
OpenAI · gpt-5-2025-08-07
2026-12-11
in 22d
OpenAI o3-pro → GPT-5.5 Pro
OpenAI · o3-pro-2025-06-10
2026-12-11
in 22d
GPT-5 pro → GPT-5.5 Pro
OpenAI · gpt-5-pro-2025-10-06
2026-12-11
in 22d
GPT-5 nano → GPT-5.4 nano
OpenAI · gpt-5-nano-2025-08-07
2026-12-11
in 22d
GPT-5 mini → GPT-5.4 mini
OpenAI · gpt-5-mini-2025-08-07
2026-12-11
in 22d
Gemini 3.1 Flash TTS Preview
Google · gemini-3.1-flash-tts-preview
2026-12-18
in 29d
Gemini 2.5 Flash Preview TTS → Gemini 3.1 Flash TTS Preview
Google · gemini-2.5-flash-preview-tts
2026-12-18
in 29d
Gemini 2.5 Pro Preview TTS → Gemini 3.1 Flash TTS Preview
Google · gemini-2.5-pro-preview-tts
2026-12-18
in 29d
GLM-4.7 → GLM-5
Z.ai (Zhipu)
2026-12-31
in 42d
GLM-4.5V
Z.ai (Zhipu)
2026-12-31
in 42d
GPT-4o → GPT-6 Luna
OpenAI · gpt-4o-2024-11-20
2027-01-14
in 56d
GPT-4.1 → GPT-6 Luna
OpenAI
2027-01-14
in 56d
Phi-4
Microsoft
2027-01-17
in 59d
Phi-4-mini-reasoning
Microsoft
2027-01-17
in 59d
Phi-4-reasoning
Microsoft
2027-01-17
in 59d
Qwen3.5 397B-A17B → Qwen3.8 Flash
Alibaba (Qwen) · qwen3.5-397b-a17b
2027-01-19
in 61d
Qwen3.5 Plus → Qwen3.8 Flash
Alibaba (Qwen) · qwen3.5-plus
2027-01-19
in 61d
Llama-3.3-70B-Instruct
Microsoft
2027-01-31
in 73d
Llama 4 Scout 17B 16E Instruct
Microsoft · llama-4-scout-17b-16e-instruct
2027-01-31
in 73d
Claude Opus 4.6
Anthropic · claude-opus-4-6
2027-02-05
in 78d
gpt-image-2
OpenAI
2027-02-10
in 83d
Command A
Microsoft · cohere-command-a
2027-02-15
in 88d
Embed v3 English
Microsoft · cohere-embed-v3-english
2027-02-15
in 88d
Embed v3 Multilingual
Microsoft · cohere-embed-v3-multilingual
2027-02-15
in 88d
Codestral 25.01
Microsoft · codestral-2501
2027-02-15
in 88d
Mistral Small 3.1
Microsoft · mistral-small-2503
2027-02-15
in 88d
Claude Sonnet 4.6
Anthropic · claude-sonnet-4-6
2027-02-17
in 90d
Claude Haiku 4.5 → Claude Haiku 5.5
Microsoft · claude-haiku-4-5
2027-02-22
in 95d
Claude Opus 4.5 → Claude Opus 5.5
Microsoft · claude-opus-4-5
2027-02-22
in 95d
Nano Banana → Nano Banana 2 Lite
Google · gemini-2.5-flash-image
2027-03-15
in 116d
GPT-5.3 Codex → GPT-6 Sol
OpenAI · gpt-5.3-codex
2027-04-01
in 133d
GPT-5.4 nano → GPT-6 Luna
OpenAI · gpt-5.4-nano
2027-04-01
in 133d
GPT-5.1 → GPT-6 Sol
OpenAI
2027-04-01
in 133d
Grok 4.20 (Non-Reasoning)
Microsoft · grok-4-20-non-reasoning
2027-04-06
in 138d
Grok 4.20 (Reasoning)
Microsoft · grok-4-20-reasoning
2027-04-06
in 138d
Claude Opus 4.7
Anthropic · claude-opus-4-7
2027-04-16
in 148d
Gemini 3.1 Flash Lite
Google · gemini-3.1-flash-lite
2027-05-07
in 169d
Grok 4.1 Fast (Non-Reasoning)
Microsoft · grok-4-1-fast-non-reasoning
2027-05-19
in 181d
Grok 4.1 Fast (Reasoning)
Microsoft · grok-4-1-fast-reasoning
2027-05-19
in 181d
Claude Opus 4.8
Anthropic · claude-opus-4-8
2027-05-28
in 190d
Claude Fable 5
Anthropic · claude-fable-5
2027-06-09
in 202d
Claude Sonnet 5
Microsoft · claude-sonnet-5
2027-06-30
in 223d
DeepSeek-V3.2 → DeepSeek-V4.1-Flash
Microsoft
2027-11-01
in 347d
DeepSeek-V3.2-Speciale → DeepSeek V4 Pro
Microsoft
2027-11-01
in 347d
Gemini Embedding 001 → Gemini Embedding 2
Google · gemini-embedding-001
2028-05-14
in 542d
GLM-5.3
Z.ai (Zhipu)
2098-12-31
in 26340d
GLM-5.3-Flash
Z.ai (Zhipu)
2098-12-31
in 26340d
GLM-5V-Turbo
Z.ai (Zhipu)
2098-12-31
in 26340d
GLM-5.3-FlashX
Z.ai (Zhipu)
2098-12-31
in 26340d
GLM-5-Turbo
Z.ai (Zhipu)
2098-12-31
in 26340d

Subscribe & explore

See what just shipped alongside what's expiring on the model radar, or subscribe to the lifecycle feed: RSS · JSON Feed · retirements only.

Frequently asked

When is Llama 4 Maverick 17B 128E Instruct FP8 retired?

Microsoft is retiring Llama 4 Maverick 17B 128E Instruct FP8 on 2026-11-19. After that date, requests to the model will fail.

How do I know if my code uses a deprecated AI model?

Run the LLM Status CLI in your repo (curl -fsSL https://llmstatus.ai/install.sh | bash, then mm scan). It finds every model id referenced in your code, env, and infrastructure and flags the ones that are deprecating or retiring, with the recommended replacement for each.

Do AI providers announce deprecations?

Yes, but inconsistently — usually in docs or changelog pages, not a single feed, and often with only a few months of notice. LLM Status tracks these announcements across providers and surfaces the shutdown date and replacement in one place.

What happens when a deprecated model is retired?

Once a model reaches its retirement (end-of-life) date, API requests to that model id return an error and stop working. Migrating to the recommended replacement before the shutdown date avoids the outage.