New dedicated endpoint models
The following models are now available for deployment on dedicated endpoints : Qwen/Qwen3-TTS-12Hz-1.7B (text-to-speech). This model was previously listed as Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice . thinkingmachines/Inkling-2-Small .
Also shipped Nov 13, 2026Together AI in November 2026 →
AI Function task support in Declarative Automation Bundles (Beta)
Databricks introduced Beta support for AI Function tasks within Declarative Automation Bundles.
Give agents only the access they need
Buildkite released token permissions and queue scopes for agent tokens, allowing teams to restrict agent capabilities and limit job acceptance to specific queues.
Many-to-many relationships in semantic views (Preview)
Snowflake Cortex AI launched a public preview for modeling many-to-many relationships between logical tables within semantic views.
BigQuery connections for MySQL, Oracle, and PostgreSQL are in preview
Google Cloud introduced BigQuery federated query connections to MySQL, Oracle, and PostgreSQL in public preview.
Upcoming deprecation of selected GitHub Copilot models in December
GitHub announced the upcoming deprecation of several Copilot models on December 14, 2026, including Claude Fable 5, Gemini 3.8 Flash, and Grok 4.6.
Aurora PostgreSQL with express configuration now supported in CloudFormation
AWS added support for Aurora PostgreSQL express configuration in CloudFormation, enabling single-property automated cluster creation and faster serverless database setup.
More from Together AI platformEvery Together AI platform release →
Post-training: Nemotron 3.5 Super VL
Together AI added post-training support for NVIDIA Nemotron 3.5 Super VL. The platform supports LoRA training on B200-SXM GPUs with context windows up to 163,840 tokens, rank up to 32, and up to three adapter sessions per resource.
Post-training: sampling-only model resources
A post-training model resource can now run generators with no trainer. To sample a saved inference checkpoint without training it further, set num_trainer_replicas=0 . You're billed only for the generator GPUs. See Sampling-only resources .
New serverless models
Together AI added thinkingmachines/Inkling-2-Small to its serverless API platform, offering a 1M token context length, NVFP4 quantization, and structured outputs.
Evaluations retired
Together AI has shut down its Evaluations service. This includes the removal of the evaluations API, associated SDK methods, CLI commands, and the console interface.
New dedicated endpoint models
The following models are now available for deployment on dedicated endpoints : nvidia/nemotron-3-asr-streaming-0.6b (speech-to-text). nvidia/nemotron-3.5-asr-streaming-0.6b (speech-to-text).
What's new:
- Rollout capacity budgets: tg endpoints rollout takes two flags that set a rollout's capacity budgets . --max-surge caps how many replicas each canary step adds to the target. --max-unavailable lets a rollout start from a DEGRADED source, or shut down source replicas to make room for the target.
Default expiry for Jig secrets
New Jig secrets now expire after 14 days unless you set a shorter lifetime with --ttl . Before a secret expires, the project's admins and the secret's creator get a daily email reminder. An expired secret is deleted, so new replicas can't read it. Existing replicas keep the value until they terminate.
New serverless models
The following models are now available on serverless : elorian-labs/glimpse-1.0 : 262,144 context length, FP8 quantization, function calling, and structured outputs. Pricing: $0.60 input / $3.00 output / $0.06 cached input (per 1M tokens).