Model decommissioning policy

更新时间:
复制 MD 格式

To optimize resources and provide users with the latest, most advanced models, Alibaba Cloud Model Studio periodically retires legacy models. This topic describes the model retirement process.

Notification process

Notification schedule

  • For snapshot models, which are identified by a specific date in their name (for example, qwen-max-2025-01-25, common for Qwen series models), we issue a sunset notice 30 days before the official sunset date.

  • For mainline models, which are the core versions of a model series, we issue a sunset notice 3 months before the official sunset date.

Notification channels

We send notifications via SMS, email, internal messages, and official website announcements.

SMS, email, and internal messages are sent only to users who have called the models scheduled for sunset in the last 3 months.

Retirement impact

  • Starting from the date of the retirement notice, the QPM (queries per minute) and TPM (tokens per minute) for retiring models will gradually decrease. For models that received a quota increase, their limits will revert to the default rate limit before this reduction begins. Throughout this period, the model API and related console features will remain fully functional.

  • Starting from the official retirement date:

    • Model inference: Model inference will be discontinued. API calls to the retired model will fail.

    • Model fine-tuning and model deployment: You will no longer be able to start new fine-tuning and deployment operations on the retired model. (For some models, these features may remain available after the retirement date. Please refer to the official retirement notice for details.) This does not affect existing trained and deployed models.

    • Console features and official documentation: Associated console features (such as Model Square and Model Discovery) and the official documentation will also be retired.

Actions

  1. Go to the Model Observation page to check if your account is using any models scheduled for sunset.

  2. If you use an affected model, test the replacement model's business performance before switching to it.

Retired models

Sunset on September 8, 2026

For more information, see the official announcement: Notice of Decommissioning for Specific Mainline Models in Model Studio.

Category

Model name

Decommission time

Replacement model

Qwen-Max

qwen3.6-max-preview

September 8, 2026, 00:00:00

qwen3.7-max

qwen3-max-preview

qwen3-max

Qwen-VL

qwen3-vl-flash

qwen3.6-flash

Qwen-Coder

qwen3-coder-plus

qwen3.7-plus

Deprecation on September 7, 2026

For details, see the official announcement: Notice of Deprecation for Legacy Speech Models on Model Studio.

Category

Model name

Deprecation date

Replacement model

Speech Synthesis

qwen-tts

September 7, 2026

cosyvoice-v3-flash

qwen-tts-realtime

Voice Cloning

qwen-voice-enrollment

voice-enrollment

Voice Design

qwen-voice-design

cosyvoice-v3.5-flash

Speech Translation

gummy-chat-v1

None

gummy-realtime-v1

Speech Recognition

paraformer-realtime-v1

paraformer-realtime-v2

paraformer-realtime-8k-v1

paraformer-realtime-8k-v2

paraformer-v1

paraformer-v2

paraformer-8k-v1

paraformer-8k-v2

paraformer-mtl-v1

paraformer-v2

Sunset: July 13, 2026

Category

Model name

Decommission time

Replacement model

Qwen-Turbo

qwen-turbo

July 13, 2026, 00:00:00

Latest models in the Qwen 3.7 and 3.6 series

qwen-turbo-realtime

Qwen-VL

qwen-vl-max

qwen-vl-plus

QwQ

qwq-plus

QVQ

qvq-max

qvq-plus

Qwen-Math

qwen-math-turbo

Qwen-Coder

qwen-coder-turbo

qwen-coder-plus

Deprecation on July 8, 2026

For details, see the announcement, Deprecation of legacy snapshot models in Model Studio.

Category

Model name

Deprecation time

Replacement model

Qwen-Max

qwen3-max-2026-01-23

July 8, 2026, 00:00:00

qwen3.7-max

qwen3-max-2025-09-23

Qwen-VL

qwen3-vl-8b-instruct

qwen3.6-flash

qwen3-vl-8b-thinking

qwen3-vl-flash-2026-01-22

qwen3-vl-flash-2025-10-15

qwen3-vl-30b-a3b-instruct

qwen3.7-plus

qwen3-vl-30b-a3b-thinking

qwen3-vl-32b-instruct

qwen3-vl-32b-thinking

qwen3-vl-235b-a22b-thinking

Qwen-Coder

qwen3-coder-next

qwen3.7-plus

qwen3-coder-30b-a3b-instruct

qwen3-coder-plus-2025-09-23

qwen3-coder-plus-2025-07-22

qwen3-coder-480b-a35b-instruct

Qwen3 Open Source

qwen3-8b

qwen3.6-flash

qwen3-14b

qwen3-30b-a3b

qwen3.7-plus

qwen3-30b-a3b-instruct-2507

qwen3-30b-a3b-thinking-2507

qwen3-32b

qwen3-235b-a22b

qwen3-vl-235b-a22b-instruct

qwen3-235b-a22b-instruct-2507

qwen3-235b-a22b-thinking-2507

qwen3-next-80b-a3b-instruct

qwen3-next-80b-a3b-thinking

third-party models

deepseek-r1-distill-qwen-7b

qwen3.7-plus

deepseek-r1-distill-qwen-14b

deepseek-r1-distill-qwen-32b

deepseek-v3

deepseek-v3.1

deepseek-v3.2

deepseek-v3.2-exp

deepseek-r1

deepseek-r1-0528

MiniMax-M2.1

glm-4.7

glm-4.6

Moonshot-Kimi-K2-Instruct

kimi-k2-thinking

Model retirement on July 6, 2026

For details, see the official announcement Notice: Retirement of Select Legacy Speech Snapshot Models on Model Studio.

Category

Model name

Retirement time

Replacement model

speech synthesis

qwen-tts-latest

July 6, 2026, 00:00:00

cosyvoice-v3-flash

qwen-tts-2025-05-22

qwen-tts-2025-04-10

qwen-tts-realtime-latest

qwen-tts-realtime-2025-07-15

Retires on May 30, 2026

For details, see the official announcement Notice on the Sunset of the GTE-RERANK model.

Category

Model name

Retirement date

Replacement model

reranking

gte-rerank

May 30, 2026, 00:00:00

qwen3-rerank

Offline

For details, see Notice on Retiring Some Historical Snapshot Models.

Category

Model name

Deprecation time

Replacement model

Qwen-Max snapshot

qwen-max-latest

May 13, 2026, 00:00:00

The latest models in the Qwen3.6 series

qwen-max-2025-01-25

qwen-max-0919

qwen-max-0428

Qwen-Turbo snapshot

qwen-turbo-latest

qwen-turbo-2025-07-15

qwen-turbo-2025-04-28

qwen-turbo-2025-02-11

qwen-turbo-2024-11-01

qwen-turbo-1101

Qwen-VL-Max snapshot

qwen-vl-max-latest

qwen-vl-max-2025-08-13

qwen-vl-max-2025-04-08

qwen-vl-max-2025-04-02

qwen-vl-max-2025-01-25

qwen-vl-max-1230

qwen-vl-max-1119

Qwen-VL-Plus snapshot

qwen-vl-plus-latest

qwen-vl-plus-2025-08-15

qwen-vl-plus-2025-07-10

qwen-vl-plus-2025-05-07

qwen-vl-plus-2025-01-25

qwen-vl-plus-0102

QwQ-Plus snapshot

qwq-plus-latest

qwq-plus-2025-03-05

QVQ-Max snapshot

qvq-max-latest

qvq-max-2025-05-15

qvq-max-2025-03-25

QVQ-Plus snapshot

qvq-plus-latest

qvq-plus-2025-05-15

Qwen-Math-Turbo snapshot

qwen-math-turbo-latest

qwen-math-turbo-0919

Qwen-Coder-Turbo snapshot

qwen-coder-turbo-latest

qwen-coder-turbo-0919

Qwen-Coder-Plus snapshot

qwen-coder-plus-latest

qwen-coder-plus-2024-11-06

Open source series snapshot

qwq-32b

qwq-32b-preview

qvq-72b-preview

qwen2.5-vl-32b-instruct

qwen2.5-vl-72b-instruct

qwen2.5-vl-7b-instruct

qwen2.5-vl-3b-instruct

qwen2.5-7b-instruct-1m

qwen2.5-14b-instruct-1m

qwen2.5-72b-instruct

qwen2.5-32b-instruct

qwen2.5-14b-instruct

qwen2.5-math-72b-instruct

qwen2.5-math-7b-instruct

qwen2.5-coder-1.5b-instruct

qwen2.5-coder-0.5b-instruct

qwen2.5-coder-14b-instruct

qwen2.5-coder-32b-instruct

qwen2.5-coder-3b-instruct

qwen2.5-coder-7b-instruct

qwen2.5-math-1.5b-instruct

qwen2.5-3b-instruct

qwen2.5-1.5b-instruct

qwen2.5-0.5b-instruct

qwen2.5-7b-instruct

qwen3-0.6b

qwen3-1.7b

qwen3-4b

Discontinued on March 30, 2026

Category

Model name

Sunset time

Replacement model

Qwen Audio

qwen-audio-asr

March 30, 2026, 00:00:00

qwen3-asr-flash

qwen-audio-asr-latest

qwen-audio-chat

qwen3-omni-flash

qwen2-audio-instruct

Qwen2 Open Source

qwen2-57b-a14b-instruct

qwen3-235b-a22b

qwen2-72b-instruct

qwen2-7b-instruct

qwen2-1.5b-instruct

qwen2-0.5b-instruct

Qwen1.5

qwen1.5-110b-chat

qwen3-235b-a22b

qwen1.5-72b-chat

qwen1.5-32b-chat

qwen1.5-14b-chat

qwen1.5-7b-chat

qwen1.5-1.8b-chat

qwen1.5-0.5b-chat

Qwen Math

qwen2.5-math-1.5b-instruct

qwen-math-plus

Qwen Coder

qwen2.5-coder-3b-instruct

qwen-coder-plus

qwen2.5-coder-1.5b-instruct

qwen2.5-coder-0.5b-instruct

Qwen-VL

qwen2-vl-72b-instruct

qwen3.5-flash

qwen2-vl-7b-instruct

qwen2-vl-2b-instruct

qwen-vl-v1

qwen-vl-chat-v1

Third-party models

baichuan2-turbo

qwen-flash

abab6.5s-chat

abab6.5g-chat

abab6.5t-chat

NLU

opennlu-v1

qwen3.5-flash

Image generation

stable-diffusion-v1.5

qwen-image-plus, z-image-turbo, wan2.6-t2i

stable-diffusion-xl

stable-diffusion-3.5-large

stable-diffusion-3.5-large-turbo

flux-dev

flux-merged

flux-schnell

Llama 4

llama-4-scout-17b-16e-instruct

qwen3.5-flash

llama-4-maverick-17b-128e-instruct

Decommissioned on January 30, 2026

For details, see the announcement, Sunsetting of Certain Historical Snapshot Models in Model Studio.

Category

Model name

Retirement date

Replacement model

Qwen-Max

qwen-max-2024-04-03

January 30, 2026, 00:00:00

qwen-max-2025-01-25

Qwen-Plus

qwen-plus-2024-11-27

qwen-plus-2025-12-01

qwen-plus-2024-11-25

qwen-plus-2024-09-19

qwen-plus-2024-08-06

qwen-plus-2024-07-23

Qwen-Turbo

qwen-turbo-2024-09-19

qwen-flash-2025-07-28

qwen-turbo-2024-06-24

Qwen-VL

qwen-vl-max-2024-10-30

qwen3-vl-plus-2025-12-19

qwen-vl-max-2024-08-09

qwen-vl-plus-2024-08-09

qwen3-vl-flash-2025-10-15

Qwen-Audio

qwen-audio-turbo-2024-12-04

qwen3-asr-flash

qwen-audio-turbo-2024-08-07

qwen-audio-asr-2024-12-04

Retired on July 30, 2025

For details, see the official announcement: [Model Studio] Notice on Retiring Legacy Models.

Category

Model name

Retirement date

Replacement model

Qwen-VL snapshot version

qwen-vl-plus-2023-12-01

July 30, 2025, 00:00:00

qwen-vl-plus

01.AI

yi-large

qwen-max, qwen-plus, qwen-flash, and others

yi-medium

yi-large-rag

yi-large-turbo

Dolly

dolly-12b-v2

Sunset: July 2, 2025

For details, see the notice Retiring legacy models on Alibaba Cloud Model Studio.

Category

Model name

Decommission time

Replacement model

Llama-text-only input

llama3.3-70b-instruct

July 2, 2025, 00:00:00

qwen-max, qwen-plus, qwen-flash, and others

llama3.2-3b-instruct

llama3.2-1b-instruct

llama3.1-405b-instruct

llama3.1-70b-instruct

llama3.1-8b-instruct

llama3-70b-instruct

llama3-8b-instruct

llama2-13b-chat-v2

llama2-7b-chat-v2

Llama-text and image input

llama3.2-90b-vision-instruct

llama3.2-11b-vision

Baichuan-open-source edition

baichuan2-13b-chat-v1

baichuan2-7b-chat-v1

baichuan-7b-v1

ChatGLM

chatglm3-6b

chatglm-6b-v2

Ziya

ziya-llama-13b-v1

BELLE

belle-llama-13b-2m-v1

ChatYuan

chatyuan-large-v2

BiLLa

billa-7b-sft-v1

anime character generation

wanx-style-cosplay-v1

No direct replacement models

image captioning

wanx-ast

creative text generation-WordArt Jinshu

wordart-surnames

AnyText image-text fusion

wanx-anytext-v1

Retired on May 8, 2025

For details, see the announcement Sunsetting Some Legacy Snapshot Models on Alibaba Cloud Model Studio.

Category

Model name

Deprecation time

Replacement model

Text generation - Qwen

qwen-max-2024-01-07

Also known as qwen-max-0107

May 8, 2025, 00:00:00

qwen-max

qwen-plus-2024-06-24

Also known as qwen-plus-0624

qwen-plus

qwen-plus-2024-02-06

Also known as qwen-plus-0206

qwen-turbo-2024-02-06

Also known as qwen-turbo-0206

qwen-turbo

qwen-vl-max-2024-02-01

Also known as qwen-vl-max-0201

qwen-vl-max

Text generation - Qwen - open-source edition

qwen-72b-chat

qwen2.5-72b-instruct

qwen-14b-chat

qwen2.5-14b-instruct

qwen-7b-chat

qwen2.5-7b-instruct

qwen-1.8b-chat

qwen2.5-1.5b-instruct

qwen-1.8b-longcontext-chat

qwen2.5-1.5b-instruct

qwen2-math-72b-instruct

qwen2.5-math-72b-instruct

qwen2-math-7b-instruct

qwen2.5-math-7b-instruct

qwen2-math-1.5b-instruct

qwen2.5-math-1.5b-instruct

Motionshop portrait video generation models

motionshop-video-detect

The 'Generate with Video Background' feature in animate-anyone-gen2 provides a similar result.

motionshop-gen3d

motionshop-synthesis

Offline as of April 22, 2024

Category

Model name

Deprecation date

Replacement model

text generation - Qwen

qwen-max-1201

April 22, 2024, 00:00:00

qwen-max