Together AI

Together AI expands AI Model Catalog with Vision, Chat, Language, Embeddings, Images, and Voice models

Together AI has added multiple new AI models across inference categories including Vision, Chat, Language, Embeddings, Images, and Voice models

Date not specified

Together AI expands inference services with new AI models

Multiple new AI models added across inference categories including Vision, Chat, Language, Embeddings, Images, and Voice services.

Date not specified

Together AI Website Downtime - December 5, 2025

Together AI's website was down for 23 minutes on December 5, 2025, potentially impacting service availability

Date not specified

Together AI Website Experienced 23-Minute Partial Downtime — Service Disruption

Together AI website was down for 23 minutes during maintenance or service disruption

Date not specified

Together AI Website Experienced Brief Downtime

Together AI website was down for 23 minutes on December 5, 2025, potentially disrupting service access

Date not specified

Together AI Platform experienced 23-minute downtime

Platform experienced a 23-minute service interruption affecting website and potentially inference services

Date not specified

Together AI Website Experienced Brief Downtime - Incident

Together AI website was down for 23 minutes on Dec 05, 2025

Date not specified

Together AI Website Downtime - 23 Minute Outage on 2025-12-05

Together AI's website was down for 23 minutes on December 5, 2025, interrupting service availability.

Date not specified

Together AI Platform experienced 23-minute downtime on Dec 05, 2025

Platform experienced a 23-minute service disruption on Dec 05, 2025, potentially affecting AI inference and model access

Date not specified

Provider Incident Detected — Critical Severity

Date not specified

Together AI Website Experienced 23-Minute Downtime — Service Interruption

Platform experienced a 23-minute service interruption affecting website availability

Date not specified

Together AI Website Experienced 23-Minute Downtime

Together AI reported a 23-minute service disruption on their website, potentially impacting platform accessibility

Date not specified

Together AI redirects Kimi-K2, DeepSeek-V3, and DeepSeek-R1 to upgraded models

Together AI has activated automatic redirects for three model versions to their upgraded counterparts, affecting API calls to the original models. Users requiring the original versions must deploy them as Dedicated Endpoints. This is a same-lineage upgrade with compatible behavior but may impact applications relying on specific model versions.

Date not specified

Together AI Python SDK v2.0 Release Candidate released with breaking changes

Together AI has released a Python SDK v2.0 Release Candidate featuring improved typing, new APIs, and redesigned error handling. While most core APIs maintain compatibility, some APIs have breaking changes in method names and parameters. Teams should review the migration guide and plan to upgrade during the 1-month RC period before general availability.

Date not specified

Together AI adds Mistral-3-14B-Instruct-2512 model to serverless offerings

Together AI has added the mistralai/Ministral-3-14B-Instruct-2512 model to their serverless model catalog. This expands available model options for users without requiring any immediate action. Users can now access this new Mistral AI instruction-tuned model through Together AI's serverless infrastructure.

Date not specified

Together AI adds GLM-4.6 and Kimi-K2-Thinking serverless models

Together AI has added two new models to their serverless offering: zai-org/GLM-4.6 and moonshotai/Kimi-K2-Thinking. This expands the available model options for users without requiring any changes to existing implementations. Users can now access these additional models through the existing serverless infrastructure.

Date not specified

Together AI launches real-time audio capabilities with WebSocket APIs

Together AI has released comprehensive real-time audio features including WebSocket APIs for low-latency text-to-speech and speech-to-text streaming. New models include Orpheus 3B and Kokoro 82M for TTS, Voxtral Mini for speech recognition, plus speaker diarization capabilities. These are additive features that enhance existing audio processing options without requiring any immediate action.

Date not specified

Together AI deprecates FLUX.1-pro and FLUX.1-Canny-pro image models

Two FLUX image models have been deprecated and are no longer available on Together AI. FLUX.1-pro calls will automatically redirect to FLUX.1.1-pro, but FLUX.1-Canny-pro has been completely removed. Users should update their integrations to use supported models.

Date not specified

Together AI launches Video Generation API with 40+ new models

Together AI has expanded into multimedia generation by introducing a new Video Generation API alongside 40+ new image and video models including OpenAI Sora 2, Google Veo 3.0, and others. The platform now offers unified text, image, and video generation through consistent APIs with production-ready serverless endpoints. This is an additive feature expansion that enhances existing capabilities without requiring immediate action from current users.

Date not specified

Together AI Batch Inference API enhanced with UI, universal models, and 30B token rate limit

Together AI significantly upgraded their Batch Inference API with a streamlined UI, support for all serverless models and private deployments, and increased rate limits from 10M to 30B tokens per user. The update also offers 50% cost reduction compared to real-time API for most models, making it more economical for high-throughput workloads.

Date not specified