Reference Source

Meta Llama 3 (70B) — Retired — shutdown February 27, 2026

For Meta Llama 3 (70B), provider is Databricks; lifecycle status is Retired; shutdown / retirement date is February 27, 2026; recommended replacement is Comparable model on the same offering, like Llama 3.2, 3.3, or 4 model of similar size; platform / deployment scope is Databricks Foundation Model APIs — provisioned throughput, recorded from its source on 2026-09-15.

Provider
Databricks our reading
Model ID / API name
Meta Llama 3 (70B) verified
Lifecycle status
Retired
Shutdown / retirement date
February 27, 2026 verified
Recommended replacement
Comparable model on the same offering, like Llama 3.2, 3.3, or 4 model of similar size verified
Platform / deployment scope
Databricks Foundation Model APIs — provisioned throughput our reading
Sourcelearn.microsoft.com
Verified
Review by
DatasetAI model deprecation and retirement dates by provider

Values marked our reading are our classification of what the source says — the source does not print them in those words. The quote below is the evidence for each one; judge it yourself.

What the source says

Meta Llama 3 (70B) | Pay-per-token: July 23, 2024 (Meta-Llama-3-70B-Instruct); December 11, 2024 (Meta-Llama-3.1-70B-Instruct) Provisioned throughput: February 27, 2026 | Pay-per-token: Meta-Llama-4-Maverick Provisioned throughput: Comparable model on the same offering, like Llama 3.2, 3.3, or 4 model of similar size. |

— learn.microsoft.com, retrieved 2026-09-15

Where each value comes from

This source states these in separate places, so each value is shown with the passage that states it.

Shutdown / retirement date

Provisioned throughput: February 27, 2026

Recommended replacement

Provisioned throughput: Comparable model on the same offering, like Llama 3.2, 3.3, or 4 model of similar size.

— all from learn.microsoft.com, retrieved 2026-09-15

Sources disagree

More than one authority states this, and they do not state the same thing. Both are reproduced with the source each came from — deciding between them is yours, not ours.

Provider

learn.microsoft.com says provider is Databricks, as of 2026-09-15.

Meta Llama 3 (70B) | Pay-per-token: July 23, 2024 (Meta-Llama-3-70B-Instruct); December 11, 2024 (Meta-Llama-3.1-70B-Instruct) Provisioned throughput: February 27, 2026 | Pay-per-token: Meta-Llama-4-Maverick Provisioned throughput: Comparable model on the same offering, like Llama 3.2, 3.3, or 4 model of similar size. |

https://learn.microsoft.com/en-us/azure/databricks/machine-learning/retired-models-policy

docs.together.ai says provider is Together AI, as of 2026-09-15.

| 2024-08-22 | `meta-llama/Meta-Llama-3-70B` | No |

https://docs.together.ai/docs/deprecations.md

Lifecycle status

learn.microsoft.com says lifecycle status is Retired, as of 2026-09-15.

Meta Llama 3 (70B) | Pay-per-token: July 23, 2024 (Meta-Llama-3-70B-Instruct); December 11, 2024 (Meta-Llama-3.1-70B-Instruct) Provisioned throughput: February 27, 2026 | Pay-per-token: Meta-Llama-4-Maverick Provisioned throughput: Comparable model on the same offering, like Llama 3.2, 3.3, or 4 model of similar size. |

https://learn.microsoft.com/en-us/azure/databricks/machine-learning/retired-models-policy

docs.together.ai says lifecycle status is Removed, as of 2026-09-15.

| 2024-08-22 | `meta-llama/Meta-Llama-3-70B` | No |

https://docs.together.ai/docs/deprecations.md

Shutdown / retirement date

learn.microsoft.com says shutdown / retirement date is February 27, 2026, as of 2026-09-15.

Meta Llama 3 (70B) | Pay-per-token: July 23, 2024 (Meta-Llama-3-70B-Instruct); December 11, 2024 (Meta-Llama-3.1-70B-Instruct) Provisioned throughput: February 27, 2026 | Pay-per-token: Meta-Llama-4-Maverick Provisioned throughput: Comparable model on the same offering, like Llama 3.2, 3.3, or 4 model of similar size. |

https://learn.microsoft.com/en-us/azure/databricks/machine-learning/retired-models-policy

docs.together.ai says shutdown / retirement date is 2024-08-22, as of 2026-09-15.

| 2024-08-22 | `meta-llama/Meta-Llama-3-70B` | No |

https://docs.together.ai/docs/deprecations.md

Platform / deployment scope

learn.microsoft.com says platform / deployment scope is Databricks Foundation Model APIs — provisioned throughput, as of 2026-09-15.

Meta Llama 3 (70B) | Pay-per-token: July 23, 2024 (Meta-Llama-3-70B-Instruct); December 11, 2024 (Meta-Llama-3.1-70B-Instruct) Provisioned throughput: February 27, 2026 | Pay-per-token: Meta-Llama-4-Maverick Provisioned throughput: Comparable model on the same offering, like Llama 3.2, 3.3, or 4 model of similar size. |

https://learn.microsoft.com/en-us/azure/databricks/machine-learning/retired-models-policy

docs.together.ai says platform / deployment scope is Together AI serverless inference, as of 2026-09-15.

| 2024-08-22 | `meta-llama/Meta-Llama-3-70B` | No |

https://docs.together.ai/docs/deprecations.md

Sources

Last verified against source: . Due for re-check by . This page as Markdown · OKF bundle · full dataset as JSON.

This one changes, and we watch it.AI model deprecation and retirement dates by provider is re-read on a schedule and every change is dated. Subscribe: Atom feed · JSON · what has changed so far.