Skip to main content

Deprecation Process

In general, each model version has a lifecycle of 3–6 months; that is, 3–6 months after launch, the model will be officially deprecated and stop serving requests. Before service termination, we will give developers a period of time to migrate to the recommended models to avoid disrupting their business.

  1. Determine the deprecation schedule for the old model version.
  2. Starting from the official notification date (i.e., the "Notification Date"), announcements will be published via this dashboard and email reminders.
  3. After the scheduled deprecation date, the model stops serving requests; existing model inference endpoints continue to function normally.

Model Deprecation Announcements

Batch 7

Scheduled Deprecation Date: 2026-07-08

Model VersionAuto-switch Model
qwen3-30b-a3b/
qwen3-30b-a3b-thinking-2507/
qwen3-32b/
qwen3-235b-a22b/
qwen3-235b-a22b-thinking-2507/
qwen3-coder-plus/
qwen3-coder-plus-2025-07-22/
qwen3-8b/
qwen3-14b/
turing/kimi-k2-thinking/

Batch 6

Scheduled Deprecation Date: 2026-07-13

Model VersionAuto-switch Model
qwen-turbo/

Batch 5

Scheduled Deprecation Date: 2026-06-15

Model VersionAuto-switch Model
turing/gemini-2.5-flash-latest/
turing/gemini-2.5-flash-lite-latest/
turing/gemini-3-flash-latest/

Batch 4

Scheduled Deprecation Date: 2026-05-13

Model VersionAuto-switch Model
qwen-max-latest/
qwen-vl-max-latest/

Batch 3

Scheduled Deprecation Date: 2026-05-11

Model VersionAuto-switch Model
bytedance/deepseek-r1/
turing/deepseek-r1-0528/
turing/deepseek-v3-0324/
turing/deepseek-v3-latest/
doubao-seed-1-8-251215/

Batch 2

Scheduled Deprecation Date: 2026-03-26 ~ 2026-03-31

Model VersionAuto-switch Model
turing/gemini-3-pro-latestturing/gemini-3.1-pro-latest
doubao-seed-1.6-thinking-250615/
bytedance/kimi-k2-250905/

Batch 1

Reference link: Volcano Engine Model Deprecation Announcement
Scheduled Deprecation Date: 2025-11-27 14:00

Model VersionAuto-switch Model
bytedance/deepseek-r1/
turing/deepseek-v3/
turing/deepseek-v3-1226/