Deprecations - as of
Groq model deprecations
Groq lists 39 deprecated or retired models on its official pages as of Oct 17, 2026: 39 retired. No shutdown is scheduled.
- Scheduled
- 0
- Deprecated
- 0
- Retired
- 39
- As of
- Oct 17, 2026
Groq deprecation policy
When Groq deprecates a model it emails affected users, updates its deprecations page with a recommended replacement and a shutdown date, and keeps the model fully working until then (sometimes auto-upgrading requests to the replacement first). After the shutdown date the model is no longer accessible and requests to its ID return errors. Production models get a clear migration path; preview models are for evaluation and may be discontinued at short notice.
Source: Groq (GroqCloud) on console.groq.com, read Oct 17, 2026.
Scheduled shutdowns
Groq lists no shutdown dated after Oct 17, 2026.
Retired
39 modelsShut down on or before Oct 17, 2026, most recent first.
All 39 retired models
| Shut down | Model | Replacement | Notes | Source |
|---|---|---|---|---|
| groq/compound-miniCompound Mini | -None listed | Groq names no replacement model for this agentic system. | Source for groq/compound-mini on console.groq.com | |
| groq/compoundCompound | -None listed | Groq names no replacement model for this agentic system. | Source for groq/compound on console.groq.com | |
| qwen/qwen3.6-27bQwen 3.6 27B | qwen/qwen3.8-27b | Applies to free and developer tiers; enterprise committed-spend customers are not affected, and the page gives no announcement date. | Source for qwen/qwen3.6-27b on console.groq.com | |
| llama-3.3-70b-versatileLlama 3.3 70B Versatile | openai/gpt-oss-120b or qwen/qwen3.6-27b | Applies to free and developer tiers; enterprise committed-spend customers are not affected. | Source for llama-3.3-70b-versatile on console.groq.com | |
| llama-3.1-8b-instantLlama 3.1 8B Instant | openai/gpt-oss-20b | Applies to free and developer tiers; enterprise committed-spend customers are not affected. | Source for llama-3.1-8b-instant on console.groq.com | |
| qwen/qwen3-32bQwen 3 32B | openai/gpt-oss-120b | Applies to free and developer tiers; enterprise committed-spend customers are not affected. | Source for qwen/qwen3-32b on console.groq.com | |
| meta-llama/llama-4-scout-17b-16e-instructLlama 4 Scout 17B 16E Instruct | openai/gpt-oss-120b or qwen/qwen3.6-27b | Applies to free and developer tiers; enterprise committed-spend customers are not affected. | Source for meta-llama/llama-4-scout-17b-16e-instruct on console.groq.com | |
| moonshotai/kimi-k2-instruct-0905Kimi K2 Instruct 0905 | openai/gpt-oss-120b | -None listed | Source for moonshotai/kimi-k2-instruct-0905 on console.groq.com | |
| meta-llama/llama-4-maverick-17b-128e-instructLlama 4 Maverick 17B 128E Instruct | openai/gpt-oss-120b | -None listed | Source for meta-llama/llama-4-maverick-17b-128e-instruct on console.groq.com | |
| meta-llama/llama-guard-4-12bLlama Guard 4 12B | openai/gpt-oss-safeguard-20b | -None listed | Source for meta-llama/llama-guard-4-12b on console.groq.com | |
| playai-tts-arabicPlayAI TTS Arabic | canopylabs/orpheus-arabic-saudi | Text-to-speech model. | Source for playai-tts-arabic on console.groq.com | |
| playai-ttsPlayAI TTS | canopylabs/orpheus-v1-english | Text-to-speech model. | Source for playai-tts on console.groq.com | |
| moonshotai/kimi-k2-instructKimi K2 Instruct | openai/gpt-oss-120b | The table names openai/gpt-oss-120b, but the text above it names moonshotai/kimi-k2-instruct-0905 as the replacement. | Source for moonshotai/kimi-k2-instruct on console.groq.com | |
| gemma2-9b-itGemma 2 9B IT | llama-3.1-8b-instant | -None listed | Source for gemma2-9b-it on console.groq.com | |
| deepseek-r1-distill-llama-70bDeepSeek R1 Distill Llama 70B | llama-3.3-70b-versatile or openai/gpt-oss-120b | -None listed | Source for deepseek-r1-distill-llama-70b on console.groq.com | |
| llama3-8b-8192Llama 3 8B 8192 | llama-3.1-8b-instant | -None listed | Source for llama3-8b-8192 on console.groq.com | |
| llama3-70b-8192Llama 3 70B 8192 | llama-3.3-70b-versatile | -None listed | Source for llama3-70b-8192 on console.groq.com | |
| distil-whisper-large-v3-enDistil Whisper Large V3 (English) | whisper-large-v3-turbo | Speech recognition model; the page gives no announcement date. | Source for distil-whisper-large-v3-en on console.groq.com | |
| mistral-saba-24bMistral Saba 24B | qwen/qwen3-32b | -None listed | Source for mistral-saba-24b on console.groq.com | |
| qwen-qwq-32bQwen QwQ 32B | qwen/qwen3-32b | -None listed | Source for qwen-qwq-32b on console.groq.com | |
| llama-guard-3-8bLlama Guard 3 8B | meta-llama/llama-guard-4-12b | -None listed | Source for llama-guard-3-8b on console.groq.com | |
| qwen-2.5-coder-32bQwen 2.5 Coder 32B | qwen-qwq-32b or openai/gpt-oss-120b | -None listed | Source for qwen-2.5-coder-32b on console.groq.com | |
| qwen-2.5-32bQwen 2.5 32B | qwen-qwq-32b or meta-llama/llama-4-scout-17b-16e-instruct | -None listed | Source for qwen-2.5-32b on console.groq.com | |
| llama-3.3-70b-specdecLlama 3.3 70B SpecDec | meta-llama/llama-4-scout-17b-16e-instruct or llama-3.3-70b-versatile | -None listed | Source for llama-3.3-70b-specdec on console.groq.com | |
| llama-3.2-90b-vision-previewLlama 3.2 90B Vision Preview | meta-llama/llama-4-scout-17b-16e-instruct | -None listed | Source for llama-3.2-90b-vision-preview on console.groq.com | |
| llama-3.2-3b-previewLlama 3.2 3B Preview | llama-3.1-8b-instant | -None listed | Source for llama-3.2-3b-preview on console.groq.com | |
| llama-3.2-1b-previewLlama 3.2 1B Preview | llama-3.1-8b-instant | -None listed | Source for llama-3.2-1b-preview on console.groq.com | |
| llama-3.2-11b-vision-previewLlama 3.2 11B Vision Preview | meta-llama/llama-4-scout-17b-16e-instruct | -None listed | Source for llama-3.2-11b-vision-preview on console.groq.com | |
| deepseek-r1-distill-qwen-32bDeepSeek R1 Distill Qwen 32B | qwen-qwq-32b | -None listed | Source for deepseek-r1-distill-qwen-32b on console.groq.com | |
| deepseek-r1-distill-llama-70b-specdecDeepSeek R1 Distill Llama 70B SpecDec | deepseek-r1-distill-llama-70b or deepseek-r1-distill-qwen-32b | Listed with a 2025-03-24 shutdown and again in the 2025-04-14 multi-model table with the same replacements. | Source for deepseek-r1-distill-llama-70b-specdec on console.groq.com | |
| mixtral-8x7b-32768Mixtral 8x7B | mistral-saba-24b or llama-3.3-70b-versatile | -None listed | Source for mixtral-8x7b-32768 on console.groq.com | |
| llama-3.1-70b-versatileLlama 3.1 70B Versatile | llama-3.3-70b-versatile | Requests were auto-upgraded to the 3.3 model from 2024-12-20, and returned errors from 2025-01-24. | Source for llama-3.1-70b-versatile on console.groq.com | |
| llama-3.1-70b-specdecLlama 3.1 70B SpecDec | llama-3.3-70b-specdec | Requests were auto-upgraded to the 3.3 model from 2024-12-20, and returned errors from 2025-01-24. | Source for llama-3.1-70b-specdec on console.groq.com | |
| llama3-groq-8b-8192-tool-use-previewLlama 3 Groq 8B Tool Use Preview | llama-3.3-70b-versatile | Users were notified by email; the page gives no announcement date. | Source for llama3-groq-8b-8192-tool-use-preview on console.groq.com | |
| llama3-groq-70b-8192-tool-use-previewLlama 3 Groq 70B Tool Use Preview | llama-3.3-70b-versatile | Users were notified by email; the page gives no announcement date. | Source for llama3-groq-70b-8192-tool-use-preview on console.groq.com | |
| gemma-7b-itGemma 7B IT | gemma2-9b-it | -None listed | Source for gemma-7b-it on console.groq.com | |
| llama-3.2-90b-text-previewLlama 3.2 90B Text Preview | llama-3.2-90b-vision-preview or llama-3.1-70b-versatile | Announced by email in November 2024 with no day given; llama-3.1-70b-versatile was recommended for text-only workloads. | Source for llama-3.2-90b-text-preview on console.groq.com | |
| llava-v1.5-7b-4096-previewLLaVA 1.5 7B Preview | llama-3.2-11b-vision-preview | The section heading says October 18, 2024 but the table gives a 10/28/24 shutdown; announced by email in September 2024 with no day given. | Source for llava-v1.5-7b-4096-preview on console.groq.com | |
| llama-3.2-11b-text-previewLlama 3.2 11B Text Preview | llama-3.2-11b-vision-preview or llama-3.1-8b-instant | The section heading says October 18, 2024 but the table gives a 10/28/24 shutdown; llama-3.1-8b-instant was recommended for text-only workloads. | Source for llama-3.2-11b-text-preview on console.groq.com |
FAQ
Which Groq models shut down next?
No Groq model has a scheduled shutdown date as of Oct 17, 2026.
How much notice does Groq give before shutting down a model?
When Groq deprecates a model it emails affected users, updates its deprecations page with a recommended replacement and a shutdown date, and keeps the model fully working until then (sometimes auto-upgrading requests to the replacement first). After the shutdown date the model is no longer accessible and requests to its ID return errors. Production models get a clear migration path; preview models are for evaluation and may be discontinued at short notice.
How many Groq models have been retired?
This page lists 39 retired Groq models as of Oct 17, 2026. The most recent is groq/compound-mini, shut down Sep 21, 2026.
Where do these Groq dates come from?
Every row is copied from Groq's official pages at console.groq.com, read on Oct 17, 2026, and links to the page it came from. Policy: https://console.groq.com/docs/deprecations
Newsletter
Know before a model shuts down
One email per week. New model releases, deprecations and pricing changes, and what they mean for what you build.