How many Groq models are retired?
39 Groq model records have status retired.
Retired Groq models
39 Groq model records have status retired, sorted by shutdown date. The stored status is retired. The dataset records these models as already shut down. Every record on this page has a shutdown date. Each row links the model page and the replacement page. The official page is https://console.groq.com/docs/deprecations. The model page records the day that page was checked. Last updated 2026-10-26, the dataset version (the newest changelog date). See all retired models. What retired means is on Retired.
| Model | Shutdown | Replacement |
|---|---|---|
| llama-3.2-11b-text-preview The replacement column says one id is for text-only workloads. | 2024-10-28 source, checked 2026-10-26 | llama-3.2-11b-vision-preview (official), llama-3.1-8b-instant (official) |
| llava-v1.5-7b-4096-preview | 2024-10-28 source, checked 2026-10-26 | llama-3.2-11b-vision-preview (official) |
| llama-3.2-90b-text-preview The replacement column says one id is for text-only workloads. | 2024-11-25 source, checked 2026-10-26 | llama-3.2-90b-vision-preview (official), llama-3.1-70b-versatile (official) |
| gemma-7b-it | 2024-12-18 source, checked 2026-10-26 | gemma2-9b-it (official) |
| llama3-groq-70b-8192-tool-use-preview One sentence names llama-3.30-70b-versatile. That id is not in a replacement column. The column names llama-3.3-70b-versatile. | 2025-01-06 source, checked 2026-10-26 | llama-3.3-70b-versatile (official) |
| llama3-groq-8b-8192-tool-use-preview One sentence names llama-3.30-70b-versatile. That id is not in a replacement column. The column names llama-3.3-70b-versatile. | 2025-01-06 source, checked 2026-10-26 | llama-3.3-70b-versatile (official) |
| llama-3.1-70b-specdec That section says requests to the model id will return errors from the shutdown date. The page says these ids automatically upgrade on 2024-12-20, before the shutdown date. | 2025-01-24 source, checked 2026-10-26 | llama-3.3-70b-specdec (official) |
| llama-3.1-70b-versatile That section says requests to the model id will return errors from the shutdown date. The page says these ids automatically upgrade on 2024-12-20, before the shutdown date. | 2025-01-24 source, checked 2026-10-26 | llama-3.3-70b-versatile (official) |
| mixtral-8x7b-32768 | 2025-03-20 source, checked 2026-10-26 | mistral-saba-24b (official), llama-3.3-70b-versatile (official) |
| deepseek-r1-distill-llama-70b-specdec Another section on this page lists shutdown date 2025-03-24 for this id. | 2025-04-14 source, checked 2026-10-26 | deepseek-r1-distill-llama-70b (official), deepseek-r1-distill-qwen-32b (official) |
| deepseek-r1-distill-qwen-32b | 2025-04-14 source, checked 2026-10-26 | qwen-qwq-32b (official) |
| llama-3.2-11b-vision-preview | 2025-04-14 source, checked 2026-10-26 | meta-llama/llama-4-scout-17b-16e-instruct (official) |
| llama-3.2-1b-preview | 2025-04-14 source, checked 2026-10-26 | llama-3.1-8b-instant (official) |
| llama-3.2-3b-preview | 2025-04-14 source, checked 2026-10-26 | llama-3.1-8b-instant (official) |
| llama-3.2-90b-vision-preview | 2025-04-14 source, checked 2026-10-26 | meta-llama/llama-4-scout-17b-16e-instruct (official) |
| llama-3.3-70b-specdec | 2025-04-14 source, checked 2026-10-26 | meta-llama/llama-4-scout-17b-16e-instruct (official), llama-3.3-70b-versatile (official) |
| qwen-2.5-32b | 2025-04-14 source, checked 2026-10-26 | qwen-qwq-32b (official), meta-llama/llama-4-scout-17b-16e-instruct (official) |
| qwen-2.5-coder-32b | 2025-04-14 source, checked 2026-10-26 | qwen-qwq-32b (official), openai/gpt-oss-120b (official) |
| llama-guard-3-8b | 2025-06-06 source, checked 2026-10-26 | meta-llama/llama-guard-4-12b (official) |
| qwen-qwq-32b | 2025-07-14 source, checked 2026-10-26 | qwen/qwen3-32b (official) |
| mistral-saba-24b | 2025-07-30 source, checked 2026-10-26 | qwen/qwen3-32b (official) |
| distil-whisper-large-v3-en | 2025-08-23 source, checked 2026-10-26 | whisper-large-v3-turbo (official) |
| llama3-70b-8192 | 2025-08-30 source, checked 2026-10-26 | llama-3.3-70b-versatile (official) |
| llama3-8b-8192 | 2025-08-30 source, checked 2026-10-26 | llama-3.1-8b-instant (official) |
| deepseek-r1-distill-llama-70b | 2025-10-02 source, checked 2026-10-26 | llama-3.3-70b-versatile (official), openai/gpt-oss-120b (official) |
| gemma2-9b-it | 2025-10-08 source, checked 2026-10-26 | llama-3.1-8b-instant (official) |
| moonshotai/kimi-k2-instruct The paragraph says in favor of moonshotai/kimi-k2-instruct-0905. The recommended replacement column is what the table prints. | 2025-10-10 source, checked 2026-10-26 | openai/gpt-oss-120b (official), moonshotai/kimi-k2-instruct-0905 (official) |
| playai-tts | 2025-12-31 source, checked 2026-10-26 | canopylabs/orpheus-v1-english (official) |
| playai-tts-arabic | 2025-12-31 source, checked 2026-10-26 | canopylabs/orpheus-arabic-saudi (official) |
| meta-llama/llama-guard-4-12b | 2026-03-05 source, checked 2026-10-26 | openai/gpt-oss-safeguard-20b (official) |
| meta-llama/llama-4-maverick-17b-128e-instruct | 2026-03-09 source, checked 2026-10-26 | openai/gpt-oss-120b (official) |
| moonshotai/kimi-k2-instruct-0905 | 2026-04-15 source, checked 2026-10-26 | openai/gpt-oss-120b (official) |
| meta-llama/llama-4-scout-17b-16e-instruct This deprecation applies to free and developer-tier usage; enterprise customers with a committed-spend contract are not affected. | 2026-07-17 source, checked 2026-10-26 | openai/gpt-oss-120b (official), qwen/qwen3.6-27b (official) |
| qwen/qwen3-32b This deprecation applies to free and developer-tier usage; enterprise customers with a committed-spend contract are not affected. | 2026-07-17 source, checked 2026-10-26 | openai/gpt-oss-120b (official) |
| llama-3.1-8b-instant This deprecation applies to free and developer-tier usage; enterprise customers with a committed-spend contract are not affected. The models page still lists this id under Production Models. The models page prints Enterprise next to this id. | 2026-08-16 source, checked 2026-10-26 | openai/gpt-oss-20b (official) |
| llama-3.3-70b-versatile This deprecation applies to free and developer-tier usage; enterprise customers with a committed-spend contract are not affected. The models page still lists this id under Production Models. The models page prints Enterprise next to this id. | 2026-08-16 source, checked 2026-10-26 | openai/gpt-oss-120b (official), qwen/qwen3.6-27b (official) |
| qwen/qwen3.6-27b This deprecation applies to free and developer-tier usage; enterprise customers with a committed-spend contract are not affected. | 2026-09-14 source, checked 2026-10-26 | qwen/qwen3.8-27b (official) |
| groq/compound That section says requests to the model id will return errors from the shutdown date. | 2026-09-21 source, checked 2026-10-26 | Not published |
| groq/compound-mini That section says requests to the model id will return errors from the shutdown date. | 2026-09-21 source, checked 2026-10-26 | Not published |
Questions
What is the latest shutdown date among retired Groq models?
The latest shutdown date on a retired model record is 2026-09-21. 2 retired model records share that date, including groq/compound (Groq).
How many retired Groq models list a replacement?
37 of 39 retired model records list a replacement. The table links each replacement page.
Where does SunsetDiff read the dates for retired Groq models?
The official page is https://console.groq.com/docs/deprecations. Each model page records the day that page was checked. Dataset version 2026-10-26.