Retiring the Models
OCI Generative AI retires its large language models (LLMs) based on each model's type and serving mode. The LLMs serve user requests in either an on-demand serving mode or a dedicated serving mode. Review the following sections to learn about deprecation and removal times and to decide what serving mode works best for you.
Terminology
- Retirement
- When a model is retired, it's no longer available for use in the Generative AI service.
- Deprecation
- When a model is deprecated it remains available in the Generative AI service, but will have a defined amount of time that it can be used before it is retired.
About Serving Modes
- On-Demand Serving Mode
-
The on-demand serving mode is available only for pretrained foundational models and has the following characteristics:
- When OCI Generative AI releases a new model version or family, you might get an overlapping period during which both versions or two families of the same model are supported until the older model version or family is retired.
- All model family and versions are not available in all supported OCI regions. See the key features in Pretrained Foundational Models in Generative AI for the available models in each region.
- Supported On-Demand Serving Mode Models
-
The following table shows the model retirment dates for the on-demand serving mode.
Model Release Date Retirement Date Replacement Model(s) cohere.command-r-plus v1.22024-06-18At least one month after the release of the 1st replacement model. Tentative cohere.command-r-16k v1.22024-06-04At least one month after the release of the 1st replacement model. Tentative meta.llama-3-70b-instruct v1.02024-06-04At least one month after the release of the 1st replacement model. Tentative cohere.command v15.62024-02-072024-07-18cohere.command-r-plus v1.2cohere.command-light v15.62024-02-072024-07-09cohere.command-r-16k v1.2cohere.embed-english-v3.02024-02-07At least 6 months after the release of the 1st replacement model. Tentative cohere.embed-multilingual-light-v3.02024-02-07At least 6 months after the release of the 1st replacement model. Tentative meta.llama-2-70b-chat2024-01-222024-07-09meta.llama-3-70b-instruct - Dedicated Serving Mode
-
The dedicated serving mode is available for custom and pretrained base models and has the following characteristics:
- Because each hosting dedicated AI cluster can only host the same version of each model, if you decide to keep using the model version that the dedicated AI cluster is already hosting and not migrate within the overlapping time period, you can request long-term support for that version.
- Existing endpoints will continue to run.
Important
If you need a dedicated serving mode model to stay alive longer than the retirement date, create a support ticket. - Supported Dedicated Serving Mode Models
-
Model Release Date Retirement Date Replacement Model(s) cohere.command-r-plus v1.22024-06-18At least 6 months after the release of the 1st replacement model. Tentative cohere.command-r-16k v1.22024-06-04At least 6 months after the release of the 1st replacement model. Tentative meta.llama-3-70b-instruct v1.02024-06-04At least 6 months after the release of the 1st replacement model. Tentative cohere.command v15.62024-02-072025-01-18cohere.command-r-plus v1.2cohere.command-light v15.62024-02-072025-01-04cohere.command-r-16k v1.2cohere.embed-english-v3.02024-02-07At least 6 months after the release of the 1st replacement model. Tentative cohere.embed-multilingual-light-v3.02024-02-07At least 6 months after the release of the 1st replacement model. Tentative meta.llama-2-70b-chat2024-01-222025-01-04meta.llama-3-70b-instruct
Deprecation times might change in the future.
- Security Vulnerabilities and Bug Fixes for Base Models
-
The Generative AI service strives to mitigate quickly against any security issues or bug fixes that are present for any of the supported base models. Check OCI release notes to learn if you need to migrate to a different version.