Together AI deprecating Qwen3-VL-8B-Instruct and others — migrate to supported models
Action Required
Applications currently using the deprecated models will experience errors and require immediate migration to a supported model to maintain functionality.
AI Impact Summary
Together AI is deprecating several models, including Qwen3-VL-8B-Instruct, Qwen3-235B-A22B-Thinking-2507, mistralai/Mixtral-8x7B-Instruct-v0.1, and zai-org/GLM-4.5-Air-FP8. These models are no longer available on serverless, impacting applications relying on them. Users should immediately assess their dependencies and migrate to supported alternatives like Qwen3-Coder-Next-FP8 or explore the newly available Gemma models.
Affected Systems
- Date
- Date not specified
- Change type
- deprecation
- Severity
- high