AI Coding Assistant
PRODUCT OVERVIEW
Together AI is a full-stack AI Native Cloud built for developers and teams running open-source models in production. It offers serverless inference APIs across chat, vision, image, audio, video, transcription, embeddings, rerank and moderation models, plus batch inference, provisioned throughput with SLAs, dedicated model and container inference, and Together Link for running open models inside existing harnesses. Beyond inference, the platform provides GPU clusters, an AI factory, sandbox developer environments and managed storage, alongside model shaping tools for post-training, fine-tuning and evaluations. Teams can start with serverless inference and move to dedicated endpoints at scale, or contact sales for custom capacity.
Sign in at api.together.ai to access the platform console and API.
Start with serverless inference for most workloads, or explore batch inference, provisioned throughput, dedicated model inference or dedicated container inference.
Browse chat, image, code, vision and video models in the model library, or open the playground to test models.
Use post-training to define RL and SFT loops, launch managed SFT and DPO fine-tuning jobs, and measure model quality with evaluations.
Provision GPU clusters, use the AI factory for custom infrastructure, build in sandbox environments and store model weights with managed storage.
Running open-source chat, vision, image, audio and video models through APIs
Processing batch inference workloads
Committing to token-based capacity with SLAs
Deploying custom models on dedicated hardware or containers
Fine-tuning and post-training open-source models
Measuring model quality with evaluations
Provisioning GPU clusters for large-scale AI workloads
Building in sandbox environments and storing model weights securely
Together AI is a full-stack AI Native Cloud for building with open-source models, covering inference, compute, model shaping and developer environments.
It offers serverless inference, batch inference, provisioned throughput with SLAs, dedicated model inference on custom hardware and dedicated container inference for custom models.
The pricing page lists Chat, Vision, Image, Audio, Video, Transcribe, Embeddings, Rerank and Moderation categories.
Serverless inference is priced per 1M tokens with separate input and output rates, and some models list cached input rates. The lowest listed input price on the pricing page is $0.04 per 1M tokens.
Together Link lets you run open models inside the harness you already use.
Yes. The platform offers managed SFT and DPO fine-tuning jobs, post-training with custom RL and SFT loops, and evaluations to measure model quality.
Together AI provides GPU clusters, an AI factory for custom infrastructure, sandbox developer environments and managed storage for model weights and data.
The pricing page invites you to contact sales or talk to an expert for help finding the best fit.
Serverless inference is billed per 1M tokens with separate input and output prices; cached input rates are listed for some models. Batch API pricing is shown separately. Provisioned throughput, dedicated inference, GPU clusters, sandbox, managed storage, post-training and fine-tuning have their own pricing sections. Contact sales for help choosing a plan.
Pay-as-you-go API access to open-source models, starting at the lowest listed input price. Most teams start here and move to dedicated endpoints at scale.
Token-based capacity with SLAs for teams that need committed inference capacity.
Dedicated model inference on custom hardware and dedicated container inference for custom models.
Reliable GPU clusters at scale, with on-demand B200s available and options for monthly reserved capacity.
Managed SFT and DPO fine-tuning jobs for top open-source models.
Define your own RL and SFT loop for model shaping.
COMMUNITY REVIEWS
No reviews yet.
Open-source model community and Model-as-a-Service platform for hosting, evaluating, fine-tuning, and deploying AI models, datasets, and applications.
