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.
Also shipped Nov 12, 2026Together AI in November 2026 →
Genie ZeroOps is available in Beta
Genie ZeroOps is now in beta, providing automated monitoring for jobs, pipelines, and tables. It detects failures and data quality issues, investigates root causes, and proposes fixes.
Replit Agent can now buy for your business and help you reach customers
Replit updated Replit Agent with capabilities to purchase business supplies using Link and Visa and manage advertising campaigns on Google Ads.
Table-level data lineage for BigQuery pipelines in Knowledge Catalog is GA
Google Cloud made table-level data lineage in Knowledge Catalog generally available for BigQuery pipelines, enabling tracking of data flow across pipelines.
BigQuery generative AI functions support Gemini 4 Hydrogen
BigQuery generative AI functions now support the Gemini 4 Hydrogen model in general availability.
Amazon Aurora and RDS now support Graviton5-based R9g database instances
AWS announced the general availability of Graviton5-based R9g database instances for Amazon Aurora and Amazon RDS. The R9g instances offer up to 30% better performance and up to 22% improved price-performance over Graviton4 instances.
[In preview] Public Preview: SQL performance monitoring for Azure SQL Managed Instance
Microsoft launched a public preview of managed SQL performance monitoring for Azure SQL Managed Instance, providing native monitoring capabilities for business-critical database workloads.
More from Together AI platformEvery Together AI platform release →
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.
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.
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 .
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 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).
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).
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.