Llama-3.2-90B-Vision-Instruct — Retired or Deprecated or Withdrawn — shutdown 2026-06-13 or 4/14/2025 or 31 March 2026 — sources disagree
For Llama-3.2-90B-Vision-Instruct, provider is Microsoft Azure / Foundry; lifecycle status is disputed: Retired per learn.microsoft.com, Deprecated per docs.sambanova.ai, Withdrawn per ibm.com; shutdown / retirement date is disputed: 2026-06-13 per learn.microsoft.com, 4/14/2025 per docs.sambanova.ai, 31 March 2026 per ibm.com; platform / deployment scope is Azure / Microsoft Foundry, recorded from its source on 2026-08-28; source re-checked 2026-10-13.
- Provider
- Microsoft Azure / Foundry our reading
- SambaNova per docs.sambanova.ai
- IBM watsonx.ai per ibm.com
- Model ID / API name
- Llama-3.2-90B-Vision-Instruct verified
- Human-readable model name
- Llama-3.2-90B-Vision-Instruct sources disagree
- Lifecycle status
- Retired sources disagree
- Shutdown / retirement date
- 2026-06-13 sources disagree
- Platform / deployment scope
- Azure / Microsoft Foundry our reading
- SambaNova SambaCloud (model remains available on SambaStack) per docs.sambanova.ai
- watsonx.ai multitenant deployment per ibm.com
- Deprecation announced
- March 31, 2025 per docs.sambanova.ai
- Recommended replacement
- Llama-4-Maverick-17B-128E-Instruct per docs.sambanova.ai
- Release date
- 25 September 2025 per ibm.com
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
| Llama-3.2-90B-Vision-Instruct | — | Retired | 2026-06-13 | — |
— learn.microsoft.com, retrieved 2026-08-28
Where each value comes from
The value above is stated away from the sentence quoted with the answer, so the passage that states it is shown here too.
Shutdown / retirement date
| Model | Version | Lifecycle | Retirement date | Replacement | | --- | --- | --- | --- | --- | | Llama-3.2-11B-Vision-Instruct | — | Retired | 2026-06-13 | — | | Llama-3.2-90B-Vision-Instruct | — | Retired | 2026-06-13 | — |
— all from learn.microsoft.com, retrieved 2026-08-28
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.
Human-readable model name
learn.microsoft.com says human-readable model name is Llama-3.2-90B-Vision-Instruct, as of 2026-08-28.
| Llama-3.2-90B-Vision-Instruct | — | Retired | 2026-06-13 | — |
https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule
ibm.com says human-readable model name is llama-3-2-90b-vision-instruct, as of 2026-09-16.
llama-3-2-90b-vision-instruct meta-llama/llama-3-2-90b-vision-instruct | 25 September 2025 | 24 November 2025 | 31 March 2026 | llama-4-maverick-17b-128e-instruct-fp8 |
https://www.ibm.com/docs/en/watsonx/saas?topic=models-foundation-model-lifecycle
Lifecycle status
learn.microsoft.com says lifecycle status is Retired, as of 2026-08-28.
| Llama-3.2-90B-Vision-Instruct | — | Retired | 2026-06-13 | — |
https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule
docs.sambanova.ai says lifecycle status is Deprecated, as of 2026-09-15.
| `Llama-3.2-90B-Vision-Instruct` | 4/14/2025 | `Llama-4-Maverick-17B-128E-Instruct` |
https://docs.sambanova.ai/docs/en/models/deprecations.md
ibm.com says lifecycle status is Withdrawn, as of 2026-09-16.
llama-3-2-90b-vision-instruct meta-llama/llama-3-2-90b-vision-instruct | 25 September 2025 | 24 November 2025 | 31 March 2026 | llama-4-maverick-17b-128e-instruct-fp8 |
https://www.ibm.com/docs/en/watsonx/saas?topic=models-foundation-model-lifecycle
Shutdown / retirement date
learn.microsoft.com says shutdown / retirement date is 2026-06-13, as of 2026-08-28.
| Llama-3.2-90B-Vision-Instruct | — | Retired | 2026-06-13 | — |
https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule
docs.sambanova.ai says shutdown / retirement date is 4/14/2025, as of 2026-09-15.
| `Llama-3.2-90B-Vision-Instruct` | 4/14/2025 | `Llama-4-Maverick-17B-128E-Instruct` |
https://docs.sambanova.ai/docs/en/models/deprecations.md
ibm.com says shutdown / retirement date is 31 March 2026, as of 2026-09-16.
llama-3-2-90b-vision-instruct meta-llama/llama-3-2-90b-vision-instruct | 25 September 2025 | 24 November 2025 | 31 March 2026 | llama-4-maverick-17b-128e-instruct-fp8 |
https://www.ibm.com/docs/en/watsonx/saas?topic=models-foundation-model-lifecycle
Where the other sources' values come from
These values were taken from a passage elsewhere on the other source's page, not from the row quoted with it above.
Lifecycle status
Models that are being deprecated and moved to **SambaStack only**:
— docs.sambanova.ai, retrieved 2026-09-15
Shutdown / retirement date
Foundation model name API model ID | Availability date | Deprecation date | Withdrawal date | Recommended alternative model | mistral-large-2512 (Dallas data center) mistralai/mistral-large-2512 | 12 December 2025 | 8 May 2026 | 12 January 2027 | mistral-large-2512 on Deploy on demand | mistral-medium-2505 (Dallas, Frankfurt, Tokyo data centers) mistralai/mistral-medium-2505 | 7 May 2025 | 8 May 2026 | 12 January 2027 | mistral-medium-2508 on Deploy on demand | granite-8b-code-instruct (Dallas, Sydney, AWS Mumbai data centers) ibm/granite-8b-code-instruct | 13 February 2025 | 8 May 2026 | 8 August 2026 | granite-8b-code-instruct on Deploy on demand | granite-guardian-3-8b (Dallas, Sydney, AWS Mumbai, AWS GovCloud data centers) ibm/granite-guardian-3-8b | 21 October 2024 | 8 May 2026 | 8 August 2026 | | llama-3-2-11b-vision-instruct (Dallas, Frankfurt, London, Tokyo, Sydney, Toronto data centers) meta-llama/llama-3-2-11b-vision-instruct | 22 January 2025 | 8 May 2026 | 8 August 2026 | llama-3-2-11b-vision-instruct on Deploy on demand | llama-3-2-90b-vision-instruct (AWS GovCloud data center) meta-llama/llama-3-2-90b-vision-instruct | 25 September 2025 | 8 May 2026 | 8 June 2026 | llama-4-maverick-17b-128e-instruct-fp8 | llama-guard-3-11b-vision (Dallas, Tokyo, Sydney data centers) meta-llama/llama-guard-3-11b-vision | 22 January 2025 | 8 May 2026 | 8 August 2026 | llama-3-2-90b-vision-instruct on Deploy on demand | allam-1-13b-instruct (Frankfurt data center) sdaia/allam-1-13b-instruct | 21 May 2024 | 8 May 2026 | 8 June 2026 | allam-1-13b-instruct on Deploy on demand | llama-3-405b-instruct meta-llama/llama-3-405b-instruct | 23 July 2024 | 24 November 2025 | 31 March 2026 | llama-4-maverick-17b-128e-instruct-fp8 | llama-3-2-90b-vision-instruct meta-llama/llama-3-2-90b-vision-instruct | 25 September 2025 | 24 November 2025 | 31 March 2026 | llama-4-maverick-17b-128e-instruct-fp8 |
— ibm.com, retrieved 2026-09-16
Sources
- learn.microsoft.comhttps://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule
- docs.sambanova.aihttps://docs.sambanova.ai/docs/en/models/deprecations.md
- docs.sambanova.aihttps://docs.sambanova.ai/docs/en/models/deprecations
- ibm.comhttps://www.ibm.com/docs/en/watsonx/saas?topic=models-foundation-model-lifecycle