Together AI

Together AI is an AI acceleration cloud delivering fast, scalable, and reliable generative-AI infrastructure. The Together API serves open-source and proprietary foundation models for chat, embeddings, vision, audio, image and video generation, fine-tuning, code execution, and dedicated GPU compute.
15 APIs 0 Features
AILLMInferenceFoundation ModelsGPUOpen Source AI
APIs

Together Chat Completions API

OpenAI-compatible chat completions across hundreds of open-source and proprietary models including Llama, Qwen, DeepSeek, GLM, Kimi, and Mistral families with streaming, tool us...
Together Completions API

Legacy text-completion endpoint for non-chat models, OpenAI-compatible.
Together Embeddings API

Generates dense vector embeddings (e.g., multilingual-e5-large-instruct, BGE) for retrieval, RAG, and semantic-search workflows.
Together Rerank API

Reranks candidate passages against a query using cross-encoder models for higher-quality retrieval and RAG.
Together Images API

Text-to-image generation across FLUX.1, FLUX.2, Nano Banana Pro, Stable Diffusion, and Dreamshaper model families.
Together Video API

Text-to-video and image-to-video generation across multiple quality and duration tiers.
Together Audio API

Text-to-speech (MiniMax Speech, Cartesia Sonic, Kokoro, Orpheus) with sub-250ms latency, and speech-to-text (Whisper Large v3, Parakeet) with 40+ language support.
Together Vision API

Multimodal vision and document understanding using models such as Qwen 3.5 (397B and 9B) and Kimi K2.5.
Together Fine-Tuning API

Supervised fine-tuning (LoRA and full) and DPO across the Together model catalog with managed training jobs and one-click deployment.
Together Files API

Upload, list, retrieve, and delete training datasets and batch input files.
Together Models API

Lists hundreds of available models with metadata, capabilities, context window, and pricing.
Together Batch API

Asynchronous batch inference with up to 50% discount over synchronous rates; fetch results when complete.
Together Code Interpreter API

Sandboxed Python execution alongside model calls for tool-using agents and code workflows.
Together Evaluations API

LLM-as-judge evaluations with automated scoring and reports for model comparisons.
Together Dedicated Endpoints API

Provision and manage dedicated GPU-backed inference endpoints (H100, B200) with hourly billing for predictable performance and isolation.
Collections

remediation.proto

OPEN
API Collection

OPEN
Together APIs

OPEN
GraphQL

Together AI GraphQL Schema

This directory contains a conceptual GraphQL schema for the Together AI platform — an AI acceleration cloud that provides fast, scalable, and reliable generative-AI infrastructu...
GRAPHQL
Pricing Plans

Together Ai Plans Pricing

6 plans
PLANS
Rate Limits

Together Ai Rate Limits

5 limits
RATE LIMITS
FinOps

Together Ai Finops

FINOPS
Event Specifications

Together AI Streaming Inference API

AsyncAPI 2.6 description of Together AI's streaming (Server-Sent Events) inference surface. Together AI exposes OpenAI-compatible HTTP endpoints that upgrade to a `text/event-st...
ASYNCAPI
Resources

👥
GitHubOrganization
Sources

opencollection: 1.0.0
info:
  name: Together APIs
  version: 2.0.0
request:
  auth:
    type: bearer
    token: '{{bearerToken}}'
items:
- info:
    name: Deployments
    type: folder
  items:
  - info:
      name: Get the list of deployments
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/deployments
    docs: Get a list of all deployments in your project
  - info:
      name: Create a new deployment
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/deployments
      body:
        type: json
        data: '{}'
    docs: Create a new deployment with specified configuration
  - info:
      name: Get a deployment by ID or name
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/deployments/:id
      params:
      - name: id
        value: ''
        type: path
    docs: Retrieve details of a specific deployment by its ID or name
  - info:
      name: Update a deployment
      type: http
    http:
      method: PATCH
      url: https://api.together.ai/v1/deployments/:id
      params:
      - name: id
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Update an existing deployment configuration
  - info:
      name: Delete a deployment
      type: http
    http:
      method: DELETE
      url: https://api.together.ai/v1/deployments/:id
      params:
      - name: id
        value: ''
        type: path
    docs: Delete an existing deployment
  - info:
      name: Get logs for a deployment
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/deployments/:id/logs
      params:
      - name: id
        value: ''
        type: path
      - name: replica_id
        value: ''
        type: query
    docs: Retrieve logs from a deployment, optionally filtered by replica ID.
- info:
    name: Secrets
    type: folder
  items:
  - info:
      name: Get the list of project secrets
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/deployments/secrets
    docs: Retrieve all secrets in your project
  - info:
      name: Create a new secret
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/deployments/secrets
      body:
        type: json
        data: '{}'
    docs: Create a new secret to store sensitive configuration values
  - info:
      name: Get a secret by ID or name
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/deployments/secrets/:id
      params:
      - name: id
        value: ''
        type: path
    docs: Retrieve details of a specific secret by its ID or name
  - info:
      name: Update a secret
      type: http
    http:
      method: PATCH
      url: https://api.together.ai/v1/deployments/secrets/:id
      params:
      - name: id
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Update an existing secret's value or metadata
  - info:
      name: Delete a secret
      type: http
    http:
      method: DELETE
      url: https://api.together.ai/v1/deployments/secrets/:id
      params:
      - name: id
        value: ''
        type: path
    docs: Delete an existing secret
- info:
    name: DeploymentsStorage
    type: folder
  items:
  - info:
      name: Download a file
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/deployments/storage/:filename
      params:
      - name: filename
        value: ''
        type: path
    docs: Download a file by redirecting to a signed URL
- info:
    name: DeploymentsVolumes
    type: folder
  items:
  - info:
      name: Get the list of project volumes
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/deployments/storage/volumes
    docs: Retrieve all volumes in your project
  - info:
      name: Create a new volume
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/deployments/storage/volumes
      body:
        type: json
        data: '{}'
    docs: Create a new volume to preload files in deployments
- info:
    name: Volumes
    type: folder
  items:
  - info:
      name: Get a volume by ID or name
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/deployments/storage/volumes/:id
      params:
      - name: id
        value: ''
        type: path
    docs: Retrieve details of a specific volume by its ID or name
  - info:
      name: Update a volume
      type: http
    http:
      method: PATCH
      url: https://api.together.ai/v1/deployments/storage/volumes/:id
      params:
      - name: id
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Update an existing volume's configuration or contents
  - info:
      name: Delete a volume
      type: http
    http:
      method: DELETE
      url: https://api.together.ai/v1/deployments/storage/volumes/:id
      params:
      - name: id
        value: ''
        type: path
    docs: Delete an existing volume
- info:
    name: Voices
    type: folder
  items:
  - info:
      name: Fetch available voices for each model
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/voices
    docs: Fetch available voices for each model
- info:
    name: Video
    type: folder
  items:
  - info:
      name: Fetch video metadata
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/videos/:id
      params:
      - name: id
        value: ''
        type: path
    docs: Fetch video metadata
  - info:
      name: Create video
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/videos
      body:
        type: json
        data: '{}'
    docs: Create a video
- info:
    name: Chat
    type: folder
  items:
  - info:
      name: Create chat completion
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/chat/completions
      body:
        type: json
        data: '{}'
    docs: Generate a model response for a given chat conversation. Supports single queries and multi-turn conversations with
      system, user, and assistant messages.
- info:
    name: Completion
    type: folder
  items:
  - info:
      name: Create completion
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/completions
      body:
        type: json
        data: '{}'
    docs: Generate text completions for a given prompt using a language, code, or image model.
- info:
    name: Embeddings
    type: folder
  items:
  - info:
      name: Create embedding
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/embeddings
      body:
        type: json
        data: '{}'
    docs: Generate vector embeddings for one or more text inputs. Returns numerical arrays representing semantic meaning,
      useful for search, classification, and retrieval.
- info:
    name: Models
    type: folder
  items:
  - info:
      name: List all models
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/models
      params:
      - name: dedicated
        value: ''
        type: query
    docs: Lists all of Together's open-source models
  - info:
      name: Upload a custom model or adapter
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/models
      body:
        type: json
        data: '{}'
    docs: Upload a custom model or adapter from Hugging Face or S3
- info:
    name: Jobs
    type: folder
  items:
  - info:
      name: Get job status
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/jobs/:jobId
      params:
      - name: jobId
        value: ''
        type: path
    docs: Get the status of a specific job
  - info:
      name: List all jobs
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/jobs
    docs: List all jobs and their statuses
- info:
    name: Images
    type: folder
  items:
  - info:
      name: Create image
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/images/generations
      body:
        type: json
        data: '{}'
    docs: Use an image model to generate an image for a given prompt.
- info:
    name: Files
    type: folder
  items:
  - info:
      name: List all files
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/files
    docs: List the metadata for all uploaded data files.
  - info:
      name: Retrieve file metadata
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/files/:id
      params:
      - name: id
        value: ''
        type: path
    docs: Retrieve the metadata for a single uploaded data file.
  - info:
      name: Delete a file
      type: http
    http:
      method: DELETE
      url: https://api.together.ai/v1/files/:id
      params:
      - name: id
        value: ''
        type: path
    docs: Delete a previously uploaded data file.
  - info:
      name: Get file contents
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/files/:id/content
      params:
      - name: id
        value: ''
        type: path
    docs: Get the contents of a single uploaded data file.
  - info:
      name: Upload a file
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/files/upload
      body:
        type: multipart-form
        data:
        - name: purpose
          type: text
          value: ''
        - name: file_name
          type: text
          value: ''
        - name: file_type
          type: text
          value: ''
        - name: file
          type: text
          value: ''
    docs: Upload a file with specified purpose, file name, and file type.
- info:
    name: Fine-tuning
    type: folder
  items:
  - info:
      name: List all jobs
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/fine-tunes
    docs: List the metadata for all fine-tuning jobs. Returns a list of FinetuneResponseTruncated objects.
  - info:
      name: Create job
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/fine-tunes
      body:
        type: json
        data: '{}'
    docs: Create a fine-tuning job with the provided model and training data.
  - info:
      name: Estimate price
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/fine-tunes/estimate-price
      body:
        type: json
        data: '{}'
    docs: Estimate the price of a fine-tuning job.
  - info:
      name: List job
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/fine-tunes/:id
      params:
      - name: id
        value: ''
        type: path
    docs: List the metadata for a single fine-tuning job.
  - info:
      name: Delete a fine-tune job
      type: http
    http:
      method: DELETE
      url: https://api.together.ai/v1/fine-tunes/:id
      params:
      - name: id
        value: ''
        type: path
      - name: force
        value: ''
        type: query
    docs: Delete a fine-tuning job.
  - info:
      name: List job events
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/fine-tunes/:id/events
      params:
      - name: id
        value: ''
        type: path
    docs: List the events for a single fine-tuning job.
  - info:
      name: List checkpoints
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/fine-tunes/:id/checkpoints
      params:
      - name: id
        value: ''
        type: path
    docs: List the checkpoints for a single fine-tuning job.
  - info:
      name: Download model
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/finetune/download
      params:
      - name: ft_id
        value: ''
        type: query
      - name: checkpoint_step
        value: ''
        type: query
      - name: checkpoint
        value: ''
        type: query
    docs: Receive a compressed fine-tuned model or checkpoint.
  - info:
      name: Cancel job
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/fine-tunes/:id/cancel
      params:
      - name: id
        value: ''
        type: path
    docs: Cancel a currently running fine-tuning job. Returns a FinetuneResponseTruncated object.
  - info:
      name: Get metrics
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/fine-tunes/:id/metrics
      params:
      - name: id
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: 'Retrieves recorded training metrics for a fine-tuning job in chronological order. All filter fields are optional
      — omit the body or send `{}` to retrieve all metrics.

      '
  - info:
      name: List supported models
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/fine-tunes/models/supported
    docs: List models supported for fine-tuning.
  - info:
      name: Get model limits
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/fine-tunes/models/limits
      params:
      - name: model_name
        value: ''
        type: query
    docs: Get model limits for a specific fine-tuning model.
- info:
    name: Rerank
    type: folder
  items:
  - info:
      name: Create a rerank request
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/rerank
      body:
        type: json
        data: '{}'
    docs: Rerank a list of documents by relevance to a query. Returns a relevance score and ordering index for each document.
- info:
    name: Audio
    type: folder
  items:
  - info:
      name: Create audio generation request
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/audio/speech
      body:
        type: json
        data: '{}'
    docs: Generate audio from input text
  - info:
      name: Real-time text-to-speech via WebSocket
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/audio/speech/websocket
      params:
      - name: model
        value: ''
        type: query
      - name: voice
        value: ''
        type: query
      - name: max_partial_length
        value: ''
        type: query
      - name: language
        value: ''
        type: query
    docs: 'Establishes a WebSocket connection for real-time text-to-speech generation. This endpoint uses WebSocket protocol
      (wss://api.together.ai/v1/audio/speech/websocket) for bidirectional streaming communication.


      **Connection Setup:**

      - Protocol: WebSocket (wss://)

      - Authentication: Pass API key as Bearer token in Authorization header

      - Parameters: Sent as query parameters (model, voice, max_partial_length, language)


      **Client Events:**

      - `tts_session.updated`: Update session parameters like voice. T'
  - info:
      name: Create audio transcription request
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/audio/transcriptions
      body:
        type: multipart-form
        data: []
    docs: Transcribes audio into text
  - info:
      name: Create audio translation request
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/audio/translations
      body:
        type: multipart-form
        data: []
    docs: Translates audio into English
  - info:
      name: Real-time audio transcription via WebSocket
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/realtime
      params:
      - name: model
        value: ''
        type: query
      - name: input_audio_format
        value: ''
        type: query
        description: Audio format specification. Currently supports 16-bit PCM at 16kHz sample rate.
    docs: "Establishes a WebSocket connection for real-time audio transcription. This endpoint uses WebSocket protocol (wss://api.together.ai/v1/realtime)\
      \ for bidirectional streaming communication.\n\n**Connection Setup:**\n- Protocol: WebSocket (wss://)\n- Authentication:\
      \ Pass API key as Bearer token in Authorization header\n- Parameters: Sent as query parameters (model, input_audio_format)\n\
      \n**Client Events:**\n- `input_audio_buffer.append`: Send audio chunks as base64-encoded data\n  ```json\n  {\n    \"\
      type\": \"in"
- info:
    name: GPUClusterService
    type: folder
  items:
  - info:
      name: List all GPU clusters.
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/compute/clusters
    docs: List all GPU clusters.
  - info:
      name: Create GPU Cluster
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/compute/clusters
      body:
        type: json
        data: '{}'
    docs: 'Create an Instant Cluster on Together''s high-performance GPU clusters.

      With features like on-demand scaling, long-lived resizable high-bandwidth shared DC-local storage,

      Kubernetes and Slurm cluster flavors, a REST API, and Terraform support,

      you can run workloads flexibly without complex infrastructure management.

      '
  - info:
      name: Get GPU cluster by cluster ID
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/compute/clusters/:cluster_id
      params:
      - name: cluster_id
        value: ''
        type: path
    docs: Retrieve information about a specific GPU cluster.
  - info:
      name: Update a GPU Cluster.
      type: http
    http:
      method: PUT
      url: https://api.together.ai/v1/compute/clusters/:cluster_id
      params:
      - name: cluster_id
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Update the configuration of an existing GPU cluster.
  - info:
      name: Delete GPU cluster by cluster ID
      type: http
    http:
      method: DELETE
      url: https://api.together.ai/v1/compute/clusters/:cluster_id
      params:
      - name: cluster_id
        value: ''
        type: path
    docs: Delete a GPU cluster by cluster ID.
- info:
    name: RegionService
    type: folder
  items:
  - info:
      name: List regions and corresponding supported driver versions
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/compute/regions
    docs: List regions and corresponding supported driver versions
- info:
    name: SharedVolumeService
    type: folder
  items:
  - info:
      name: List all shared volumes.
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/compute/clusters/storage/volumes
    docs: List all shared volumes.
  - info:
      name: Create a shared volume.
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/compute/clusters/storage/volumes
      body:
        type: json
        data: '{}'
    docs: 'Instant Clusters supports long-lived, resizable in-DC shared storage with user data persistence.

      You can dynamically create and attach volumes to your cluster at cluster creation time, and resize as your data grows.

      All shared storage is backed by multi-NIC bare metal paths, ensuring high-throughput and low-latency performance for
      shared storage.

      '
  - info:
      name: Update a shared volume.
      type: http
    http:
      method: PUT
      url: https://api.together.ai/v1/compute/clusters/storage/volumes
      body:
        type: json
        data: '{}'
    docs: 'Update the configuration of an existing shared volume.

      '
  - info:
      name: Get shared volume by volume Id.
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/compute/clusters/storage/volumes/:volume_id
      params:
      - name: volume_id
        value: ''
        type: path
    docs: Retrieve information about a specific shared volume.
  - info:
      name: Delete shared volume by volume id.
      type: http
    http:
      method: DELETE
      url: https://api.together.ai/v1/compute/clusters/storage/volumes/:volume_id
      params:
      - name: volume_id
        value: ''
        type: path
    docs: 'Delete a shared volume. Note that if this volume is attached to a cluster, deleting will fail.

      '
- info:
    name: endpoints
    type: folder
  items:
  - info:
      name: List all available availability zones.
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/clusters/availability-zones
    docs: List all available availability zones.
- info:
    name: Endpoints
    type: folder
  items:
  - info:
      name: List all endpoints, can be filtered by type
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/endpoints
      params:
      - name: type
        value: ''
        type: query
      - name: usage_type
        value: ''
        type: query
      - name: mine
        value: ''
        type: query
    docs: Returns a list of all endpoints associated with your account. You can filter the results by type (dedicated or serverless).
  - info:
      name: Create a dedicated endpoint, it will start automatically
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/endpoints
      body:
        type: json
        data: '{}'
    docs: Creates a new dedicated endpoint for serving models. The endpoint will automatically start after creation. You can
      deploy any supported model on hardware configurations that meet the model's requirements.
  - info:
      name: Get endpoint by ID
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/endpoints/:endpointId
      params:
      - name: endpointId
        value: ''
        type: path
    docs: Retrieves details about a specific endpoint, including its current state, configuration, and scaling settings.
  - info:
      name: Update endpoint, this can also be used to start or stop a dedicated endpoint
      type: http
    http:
      method: PATCH
      url: https://api.together.ai/v1/endpoints/:endpointId
      params:
      - name: endpointId
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Updates an existing endpoint's configuration. You can modify the display name, autoscaling settings, or change the
      endpoint's state (start/stop).
  - info:
      name: Delete endpoint
      type: http
    http:
      method: DELETE
      url: https://api.together.ai/v1/endpoints/:endpointId
      params:
      - name: endpointId
        value: ''
        type: path
    docs: Permanently deletes an endpoint. This action cannot be undone.
- info:
    name: Hardware
    type: folder
  items:
  - info:
      name: List available hardware configurations
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/hardware
      params:
      - name: model
        value: ''
        type: query
    docs: 'Returns a list of available hardware configurations for deploying models. When a model parameter is provided, it
      returns only hardware configurations compatible with that model, including their current availability status.

      '
- info:
    name: Code Interpreter
    type: folder
  items:
  - info:
      name: tci/execute
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/tci/execute
      body:
        type: json
        data: '{}'
    docs: 'Executes the given code snippet and returns the output. Without a session_id, a new session will be created to
      run the code. If you do pass in a valid session_id, the code will be run in that session. This is useful for running
      multiple code snippets in the same environment, because dependencies and similar things are persisted

      between calls to the same session.

      '
  - info:
      name: sessions/list
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/tci/sessions
    docs: 'Lists all your currently active sessions.

      '
- info:
    name: Batches
    type: folder
  items:
  - info:
      name: List batch jobs
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/batches
      auth:
        type: bearer
        token: '{{bearerToken}}'
    docs: List all batch jobs for the authenticated user
  - info:
      name: Create a batch job
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/batches
      body:
        type: json
        data: '{}'
      auth:
        type: bearer
        token: '{{bearerToken}}'
    docs: Create a new batch job with the given input file and endpoint
  - info:
      name: Get a batch job
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/batches/:id
      params:
      - name: id
        value: ''
        type: path
      auth:
        type: bearer
        token: '{{bearerToken}}'
    docs: Get details of a batch job by ID
  - info:
      name: Cancel a batch job
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/batches/:id/cancel
      params:
      - name: id
        value: ''
        type: path
      auth:
        type: bearer
        token: '{{bearerToken}}'
    docs: Cancel a batch job by ID
- info:
    name: evaluation
    type: folder
  items:
  - info:
      name: Get all evaluation jobs
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/evaluation
      params:
      - name: status
        value: ''
        type: query
      - name: limit
        value: ''
        type: query
    docs: Get all evaluation jobs
  - info:
      name: Create an evaluation job
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/evaluation
      body:
        type: json
        data: '{}'
    docs: Create an evaluation job
  - info:
      name: Get model list
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/evaluation/model-list
      params:
      - name: model_source
        value: ''
        type: query
    docs: Get model list
  - info:
      name: Get evaluation job details
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/evaluation/:id
      params:
      - name: id
        value: ''
        type: path
    docs: Get evaluation job details
  - info:
      name: Get evaluation job status and results
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/evaluation/:id/status
      params:
      - name: id
        value: ''
        type: path
    docs: Get evaluation job status and results
- info:
    name: Queue
    type: folder
  items:
  - info:
      name: Cancel a queued job
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/queue/cancel
      body:
        type: json
        data: '{}'
    docs: 'Cancel a pending job. Only jobs in pending status can be canceled.

      Running jobs cannot be stopped. Returns the job status after the

      attempt. If the job is not pending, returns 409 with the current status

      unchanged.

      '
  - info:
      name: Get queue metrics
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/queue/metrics
      params:
      - name: model
        value: ''
        type: query
    docs: Get the current queue statistics for a model, including pending and running job counts.
  - info:
      name: Get job status
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/queue/status
      params:
      - name: request_id
        value: ''
        type: query
      - name: model
        value: ''
        type: query
    docs: Poll the current status of a previously submitted job. Provide the request_id and model as query parameters.
  - info:
      name: Submit a queued job
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/queue/submit
      body:
        type: json
        data: '{}'
    docs: 'Submit a new job to the queue for asynchronous processing. Jobs are

      processed in strict priority order (higher priority first, FIFO within

      the same priority). Returns a request ID that can be used to poll status

      or cancel the job.

      '
- info:
    name: RL
    type: folder
  items:
  - info:
      name: List training sessions
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/rl/training-sessions
      params:
      - name: status
        value: ''
        type: query
      - name: limit
        value: ''
        type: query
      - name: after
        value: ''
        type: query
    docs: Lists all training sessions.
  - info:
      name: Create training session
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/rl/training-sessions
      body:
        type: json
        data: '{}'
    docs: Creates a training session and returns its details.
  - info:
      name: Get training session
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/rl/training-sessions/:session_id
      params:
      - name: session_id
        value: ''
        type: path
    docs: Gets a training session by its ID and returns its details.
  - info:
      name: Stop training session
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/rl/training-sessions/:session_id/stop
      params:
      - name: session_id
        value: ''
        type: path
    docs: Stops a training session.
  - info:
      name: Get forward-backward operation
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/rl/training-sessions/:session_id/operations/forward-backward/:operation_id
      params:
      - name: session_id
        value: ''
        type: path
      - name: operation_id
        value: ''
        type: path
    docs: Retrieves the current status and result of a forward-backward operation.
  - info:
      name: Get optim-step operation
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/rl/training-sessions/:session_id/operations/optim-step/:operation_id
      params:
      - name: session_id
        value: ''
        type: path
      - name: operation_id
        value: ''
        type: path
    docs: Retrieves the current status and result of an optim-step operation.
  - info:
      name: Get sample operation
      type: http
    http:
      method: GET
      url: https://api.together.ai/v1/rl/training-sessions/:session_id/operations/sample/:operation_id
      params:
      - name: session_id
        value: ''
        type: path
      - name: operation_id
        value: ''
        type: path
    docs: Retrieves the current status and result of a sample operation.
  - info:
      name: Forward-backward pass
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/rl/training-sessions/:session_id/operations/forward-backward
      params:
      - name: session_id
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Submits a forward-backward pass operation that will asynchronously compute gradients via backpropagation.
  - info:
      name: Optimizer step
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/rl/training-sessions/:session_id/operations/optim-step
      params:
      - name: session_id
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Submits an optimizer step operation that will asynchronously apply accumulated gradients to update model parameters.
  - info:
      name: Sample
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/rl/training-sessions/:session_id/operations/sample
      params:
      - name: session_id
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Submits a sample operation that will asynchronously generate text completions with logprobs.
  - info:
      name: Create inference checkpoint
      type: http
    http:
      method: POST
      url: https://api.together.ai/v1/rl/training-sessions/:session_id/operations/inference-checkpoint
      params:
      - name: session_id
        value: ''
        type: path
    docs: Submits an operation that will asynchronously save the current LoRA adapter as an inference checkpoint and upload
      it to object storage.
  - info:
      name: Get inference checkpoint operation
      type: http
    http:
      method: GET
      url:

# --- truncated at 32 KB (33 KB total) ---
# Full source: https://raw.githubusercontent.com/api-evangelist/together-ai/refs/heads/main/apis.yml