---
title: "List supported models"
method: GET
path: "/supported-models"
tags: ["SupportedModelsService"]
---

# List supported models

`GET /supported-models`

Lists Together-hosted base models that can be deployed for dedicated inference, together with their capabilities and certified deployment profiles.

## Query parameters

- `modality` 'MODALITY_TEXT' | 'MODALITY_IMAGE' | 'MODALITY_AUDIO' | 'MODALITY_VIDEO' — Filter models by input modality.
- `product` 'PRODUCT_SERVERLESS' | 'PRODUCT_DEDICATED' | 'PRODUCT_FINE_TUNING' — Filter models by product surface.
- `search` string — Case-insensitive search across model IDs, names, and descriptions.
- `limit` integer — Maximum number of models to return.
- `after` string — Cursor from a previous supported-model list response.

## Response `200`

OK

- DEListSupportedModelsResponse — Response containing supported models and pagination metadata.
  - `data` DESupportedModel[], required — Supported models in the catalog.
    - `id` string, required — Unique ID of the deployable Together-hosted base model.
    - `name` string, required — Catalog-controlled HF model ID used for inference.
    - `displayName` string, required — Catalog-controlled human-readable model name.
    - `description` string — Human-readable model description.
    - `inputModalities` string[], required — Input modalities supported by the model.
    - `outputModalities` string[], required — Output modalities produced by the model.
    - `products` string[], required — Product surfaces where the model is offered.
    - `features` string[] — Advanced features exposed by the model.
    - `capabilities` string[], required — High-level tasks the model supports.
    - `architecture` string — Model architecture from the underlying weights metadata.
    - `contextLength` string — Maximum context length from the underlying weights metadata.
    - `publisher` string, required — Organization or publisher associated with the model.
    - `status` 'SUPPORTED_MODEL_STATUS_RECOMMENDED' | 'SUPPORTED_MODEL_STATUS_SUPPORTED' | 'SUPPORTED_MODEL_STATUS_DEPRECATED' | 'SUPPORTED_MODEL_STATUS_HIDDEN', required — Catalog recommendation status for the model.
    - `tags` string[] — Searchable catalog tags for the model.
    - `inputFormat` string — Preferred input format for the model.
    - `outputFormat` string — Preferred output format for the model.
    - `serverlessEndpoint` string — Serverless endpoint name for inference, if available.
    - `familyId` string — Model family identifier for related catalog entries.
    - `displayType` string, required — UI-facing model type badge, such as chat, language, code, image, embedding, rerank, moderation, audio, video, or transcribe.
    - `baseModelId` string, required — Bare model ID for the architecture's base model; empty when no base model is linked.
    - `baseModel` string, required — Resource name for the base model as `projects/{projectId}/models/{modelId}`; empty when unresolved.
    - `deploymentProfiles` DESupportedModelDeploymentProfile[], required — Certified deployment profiles available for the model.
      - `profileId` string, required — Stable profile identifier, usually the certified config id.
      - `certifiedConfigRevisionId` string, required — Certified configuration revision identifier.
      - `certifiedModelRevisionId` string, required — Certified model weight revision identifier, if available.
      - `gpuType` string, required — GPU instance type for the profile.
      - `gpuCount` integer, required — Number of GPUs required by the profile.
      - `quantization` string, required — Quantization method for the profile, if available.
      - `tensorParallelSize` integer — Deprecated. Use `parallelism`. Legacy tensor-parallel shard count for the profile.
      - `performanceBenchmarks` DESupportedModelPerformanceBenchmarks, required — Performance benchmark metrics for a supported model profile.
        - `decodingSpeedTps` number, double — Decoding throughput in tokens per second.
        - `timeToFirstTokenMs` integer — Time to first token in milliseconds.
        - `maxContextLength` string — Maximum context length supported by the profile.
      - `config` string, required — Certified config revision in the form `projects/{projectId}/configs/{configRevisionId}`. Omitted when the profile does not pin a config.
      - `model` string, required — Deployable model resource in the form `projects/{projectId}/models/{modelId}[/revisions/{revisionId}]`. Omitted when the profile does not pin model weights.
      - `parallelism` string, required — Free-form parallelism spec for the profile, such as TP8, TP4, EP, or PD; supersedes tensor_parallel_size.
      - `modelName` string, required — Fully-qualified deploy model name in the form `{projectSlug}/{modelName}`, such as `Qwen/Qwen3.5-9B-FP8`; empty when no public model is linked.
    - `createdAt` string, date-time, required — Timestamp when the catalog entry was created.
    - `updatedAt` string, date-time, required — Timestamp when the catalog entry was last updated.
  - `next_cursor` string — Cursor for the next page. Null if there are no more results.
  - `object` 'list', required — Object type. Always `list`.

## Other responses

- `default` — Default error response

---

[API](https://skmtc.net/together/apis/together-apis.md) · [All operations](https://skmtc.net/together/apis/together-apis/llms.txt) · [OpenAPI document](https://skmtc-service-staging.skmtc.workers.dev/v1/apis/together/together-apis/revisions/468edbdc879c/schema)