Together AI deprecates four models on serverless
Together AI has deprecated openai/gpt-oss-20b, google/gemma-4-31B-it, thinkingmachines/Inkling-Small, and intfloat/multilingual-e5-large-instruct on serverless.
Models:openai/gpt-oss-20bQwen/Qwen3.5-9Bgoogle/gemma-4-31B-itzai-org/GLM-5.3-Flashthinkingmachines/Inkling-Smallintfloat/multilingual-e5-large-instruct
What changed
Together AI deprecated four models and removed them from serverless availability. The models are openai/gpt-oss-20b, google/gemma-4-31B-it, thinkingmachines/Inkling-Small, and intfloat/multilingual-e5-large-instruct. The first three are still supported by on demand dedicated endpoints. Together AI recommends Qwen/Qwen3.5-9B and zai-org/GLM-5.3-Flash as replacements.
What it means for Zabio customers
Zabio does not offer Together AI models today. Developers using these models directly through Together AI must migrate to the recommended replacements or use dedicated endpoints.
What Zabio does
Catalog change under review.
Source
The following models have been deprecated and are no longer available on serverless:
Original announcement: Together AI changelog
Published automatically from the official announcement. Reviewed by a person within the day.
Source: Together AI changelog (2026-09-15)