2026-09-29 06:18 UTC · groq/deprecations · live page
Summary by fallback. It may be wrong; the diff below is what actually changed. Raw: a73270cdbb-groq-deprecations.diff · commit a73270cdbb
diff --git a/snapshots/groq/deprecations.md b/snapshots/groq/deprecations.mdindex e1f60e4..b057167 100644--- a/snapshots/groq/deprecations.md+++ b/snapshots/groq/deprecations.md@@ -39,4 +39,17 @@ When a model is marked for deprecation, we follow this standardized process: - Consider designing your systems to be model-agnostic where possible +On August 24, 2026, we announced the deprecation of `groq/compound` and `groq/compound-mini`. Both systems will be decommissioned on September 21, 2026. Beginning on that date, requests to these model IDs will return errors. Historical overviews are available for Compound and Compound Mini, along with the Compound changelog.++| Deprecated Model | Shutdown Date | Recommended Replacement Model ID |+|---|---|---|+| `groq/compound` | 09/21/26 | — |+| `groq/compound-mini` | 09/21/26 | — |++In line with our commitment to bringing you cutting-edge models, we announced the deprecation of `qwen/qwen3.6-27b` in favor of `qwen/qwen3.8-27b`. Qwen 3.8 27B is the direct successor: a 27B multimodal model with the same 131K context window, thinking and instruct modes, tunable reasoning effort, tool use, and JSON mode. This deprecation applies to free and developer-tier usage; enterprise customers with a committed-spend contract are not affected.++| Deprecated Model | Shutdown Date | Recommended Replacement Model ID |+|---|---|---|+| `qwen/qwen3.6-27b` | 09/14/26 | `qwen/qwen3.8-27b` |+ In line with our commitment to bringing you cutting-edge models, on June 17, 2026, we emailed users to announce the deprecation of `llama-3.1-8b-instant` and `llama-3.3-70b-versatile`. We recommend migrating to `openai/gpt-oss-20b` (for Llama 3.1 8B Instant) and `openai/gpt-oss-120b` or `qwen/qwen3.6-27b` (for Llama 3.3 70B Versatile), which deliver exceptional performance with faster inference. This deprecation applies to free and developer-tier usage; enterprise customers with a committed-spend contract are not affected.