together 2.0.0a16__tar.gz → 2.0.0a18__tar.gz

together-2.0.0a18/.release-please-manifest.json

@@ -0,0 +1,3 @@
+{
+  ".": "2.0.0-alpha.18"
+}

{together-2.0.0a16 → together-2.0.0a18}/CHANGELOG.md

@@ -1,5 +1,55 @@
 # Changelog
 
+## 2.0.0-alpha.18 (2026-01-31)
+
+Full Changelog: [v2.0.0-alpha.17...v2.0.0-alpha.18](https://github.com/togethercomputer/together-py/compare/v2.0.0-alpha.17...v2.0.0-alpha.18)
+
+### Features
+
+* Add chat completion support for reasoning.enabled ([7fba349](https://github.com/togethercomputer/together-py/commit/7fba349c1bde34ace5ca38c0728479f054518e04))
+* **api:** Add API for listing deployments ([298d447](https://github.com/togethercomputer/together-py/commit/298d4475208fb24e1373fa7449c0716768d5501d))
+* **api:** Add beta sdks for jig features ([c3ac883](https://github.com/togethercomputer/together-py/commit/c3ac883f8ffe62d2c1b660e3a62fc82891e63dac))
+* **api:** Move jobs apis to nest under model uploads per their use case ([368c003](https://github.com/togethercomputer/together-py/commit/368c003ec61e8cea2ec995ba9889d3cf7ac21328))
+* **api:** Move queue out of jig namespace ([ebc1498](https://github.com/togethercomputer/together-py/commit/ebc1498125e7b95f2726cfb37a4324aadcf40f8d))
+* **api:** Update Jig types and add retrieve_logs api to jig ([76a3c1a](https://github.com/togethercomputer/together-py/commit/76a3c1abccf8bb1efaebd192a68fe31785b568cf))
+* **cli:** enhance hardware command to display availability status ([#223](https://github.com/togethercomputer/together-py/issues/223)) ([e7cf8a3](https://github.com/togethercomputer/together-py/commit/e7cf8a30c6a1d46f10ea5049f51e969be6f8418f))
+* **client:** add custom JSON encoder for extended type support ([036a0ea](https://github.com/togethercomputer/together-py/commit/036a0eaba07c3cca56e9dbec965e1cce2f8b848a))
+* internal: Update to new cluster api routing ([54af0bd](https://github.com/togethercomputer/together-py/commit/54af0bdf4076e5ff02f0475547a4b1d2f7393c46))
+* move byoc features under /deployments route ([d70e2e9](https://github.com/togethercomputer/together-py/commit/d70e2e9c3c12997935a34cb4a63a26e315e0c799))
+
+
+### Bug Fixes
+
+* Avoid crashing when uploading a model which weights already exist ([#226](https://github.com/togethercomputer/together-py/issues/226)) ([5a9095c](https://github.com/togethercomputer/together-py/commit/5a9095c91da5de8333e366e88f6960b428177d07))
+
+
+### Chores
+
+* **api:** Improve type names for jig volumes and logs ([b256c61](https://github.com/togethercomputer/together-py/commit/b256c61f601d67a0abee1dbc44e9851d52014d0d))
+* **api:** Move Queue SDK methods into Jig namespace ([43179e1](https://github.com/togethercomputer/together-py/commit/43179e1ee8451cc3bcb8430c2bc35148309b82db))
+* **api:** Rename jig queue apis ([4b466d6](https://github.com/togethercomputer/together-py/commit/4b466d6079071fd840880337b20c4d34a3a5cf45))
+* **ci:** upgrade `actions/github-script` ([c8668a1](https://github.com/togethercomputer/together-py/commit/c8668a1424fbfcd37d893d414d4dcd0307f5aac4))
+
+
+### Documentation
+
+* **axle-queue:** added axle-queue endpoints ([4e60b09](https://github.com/togethercomputer/together-py/commit/4e60b096128cbfef4769195a9b00cf0236e5052c))
+
+## 2.0.0-alpha.17 (2026-01-21)
+
+Full Changelog: [v2.0.0-alpha.16...v2.0.0-alpha.17](https://github.com/togethercomputer/together-py/compare/v2.0.0-alpha.16...v2.0.0-alpha.17)
+
+### Features
+
+* **cli:** add b200 and h200 GPU options for endpoint creation ([#218](https://github.com/togethercomputer/together-py/issues/218)) ([b514912](https://github.com/togethercomputer/together-py/commit/b514912a281922fefbf8a9f62b936ed1de243718))
+* Improve usage of models list cli command ([#216](https://github.com/togethercomputer/together-py/issues/216)) ([430e6c1](https://github.com/togethercomputer/together-py/commit/430e6c1e030749be474f020b677d91014ba4482c))
+
+
+### Chores
+
+* Deprecate CLI usage for endpoints create flag --no-promopt-cache ([#219](https://github.com/togethercomputer/together-py/issues/219)) ([55e9700](https://github.com/togethercomputer/together-py/commit/55e9700187b42f8baff6f567a3a657b46577ed88))
+* Mark disable_prompt_cache as deprecated for endpoint creation ([6a629b2](https://github.com/togethercomputer/together-py/commit/6a629b29e53b4374503d30ca75456184ef313b67))
+
 ## 2.0.0-alpha.16 (2026-01-18)
 
 Full Changelog: [v2.0.0-alpha.15...v2.0.0-alpha.16](https://github.com/togethercomputer/together-py/compare/v2.0.0-alpha.15...v2.0.0-alpha.16)

{together-2.0.0a16 → together-2.0.0a18}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: together
-Version: 2.0.0a16
+Version: 2.0.0a18
 Summary: The official Python library for the together API
 Project-URL: Homepage, https://github.com/togethercomputer/together-py
 Project-URL: Repository, https://github.com/togethercomputer/together-py
@@ -235,20 +235,17 @@ from together import Together
 
 client = Together()
 
-cluster = client.beta.clusters.create(
-    billing_type="RESERVED",
-    cluster_name="cluster_name",
-    driver_version="CUDA_12_5_555",
-    gpu_type="H100_SXM",
-    num_gpus=0,
-    region="us-central-8",
-    shared_volume={
-        "region": "region",
-        "size_tib": 0,
-        "volume_name": "volume_name",
-    },
+chat_completion = client.chat.completions.create(
+    messages=[
+        {
+            "content": "content",
+            "role": "system",
+        }
+    ],
+    model="meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo",
+    reasoning={},
 )
-print(cluster.shared_volume)
+print(chat_completion.reasoning)
 ```
 
 The async client uses the exact same interface. If you pass a [`PathLike`](https://docs.python.org/3/library/os.html#os.PathLike) instance, the file contents will be read asynchronously automatically.

{together-2.0.0a16 → together-2.0.0a18}/README.md

@@ -186,20 +186,17 @@ from together import Together
 
 client = Together()
 
-cluster = client.beta.clusters.create(
-    billing_type="RESERVED",
-    cluster_name="cluster_name",
-    driver_version="CUDA_12_5_555",
-    gpu_type="H100_SXM",
-    num_gpus=0,
-    region="us-central-8",
-    shared_volume={
-        "region": "region",
-        "size_tib": 0,
-        "volume_name": "volume_name",
-    },
+chat_completion = client.chat.completions.create(
+    messages=[
+        {
+            "content": "content",
+            "role": "system",
+        }
+    ],
+    model="meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo",
+    reasoning={},
 )
-print(cluster.shared_volume)
+print(chat_completion.reasoning)
 ```
 
 The async client uses the exact same interface. If you pass a [`PathLike`](https://docs.python.org/3/library/os.html#os.PathLike) instance, the file contents will be read asynchronously automatically.

{together-2.0.0a16 → together-2.0.0a18}/api.md

@@ -1,5 +1,74 @@
 # Beta
 
+## Jig
+
+Types:
+
+```python
+from together.types.beta import Deployment, DeploymentLogs, JigListResponse
+```
+
+Methods:
+
+- <code title="get /deployments/{id}">client.beta.jig.<a href="./src/together/resources/beta/jig/jig.py">retrieve</a>(id) -> <a href="./src/together/types/beta/deployment.py">Deployment</a></code>
+- <code title="patch /deployments/{id}">client.beta.jig.<a href="./src/together/resources/beta/jig/jig.py">update</a>(id, \*\*<a href="src/together/types/beta/jig_update_params.py">params</a>) -> <a href="./src/together/types/beta/deployment.py">Deployment</a></code>
+- <code title="get /deployments">client.beta.jig.<a href="./src/together/resources/beta/jig/jig.py">list</a>() -> <a href="./src/together/types/beta/jig_list_response.py">JigListResponse</a></code>
+- <code title="post /deployments">client.beta.jig.<a href="./src/together/resources/beta/jig/jig.py">deploy</a>(\*\*<a href="src/together/types/beta/jig_deploy_params.py">params</a>) -> <a href="./src/together/types/beta/deployment.py">Deployment</a></code>
+- <code title="delete /deployments/{id}">client.beta.jig.<a href="./src/together/resources/beta/jig/jig.py">destroy</a>(id) -> object</code>
+- <code title="get /deployments/{id}/logs">client.beta.jig.<a href="./src/together/resources/beta/jig/jig.py">retrieve_logs</a>(id, \*\*<a href="src/together/types/beta/jig_retrieve_logs_params.py">params</a>) -> <a href="./src/together/types/beta/deployment_logs.py">DeploymentLogs</a></code>
+
+### Queue
+
+Types:
+
+```python
+from together.types.beta.jig import (
+    QueueRetrieveResponse,
+    QueueCancelResponse,
+    QueueMetricsResponse,
+    QueueSubmitResponse,
+)
+```
+
+Methods:
+
+- <code title="get /queue/status">client.beta.jig.queue.<a href="./src/together/resources/beta/jig/queue.py">retrieve</a>(\*\*<a href="src/together/types/beta/jig/queue_retrieve_params.py">params</a>) -> <a href="./src/together/types/beta/jig/queue_retrieve_response.py">QueueRetrieveResponse</a></code>
+- <code title="post /queue/cancel">client.beta.jig.queue.<a href="./src/together/resources/beta/jig/queue.py">cancel</a>(\*\*<a href="src/together/types/beta/jig/queue_cancel_params.py">params</a>) -> <a href="./src/together/types/beta/jig/queue_cancel_response.py">QueueCancelResponse</a></code>
+- <code title="get /queue/metrics">client.beta.jig.queue.<a href="./src/together/resources/beta/jig/queue.py">metrics</a>(\*\*<a href="src/together/types/beta/jig/queue_metrics_params.py">params</a>) -> <a href="./src/together/types/beta/jig/queue_metrics_response.py">QueueMetricsResponse</a></code>
+- <code title="post /queue/submit">client.beta.jig.queue.<a href="./src/together/resources/beta/jig/queue.py">submit</a>(\*\*<a href="src/together/types/beta/jig/queue_submit_params.py">params</a>) -> <a href="./src/together/types/beta/jig/queue_submit_response.py">QueueSubmitResponse</a></code>
+
+### Volumes
+
+Types:
+
+```python
+from together.types.beta.jig import Volume, VolumeListResponse
+```
+
+Methods:
+
+- <code title="post /deployments/storage/volumes">client.beta.jig.volumes.<a href="./src/together/resources/beta/jig/volumes.py">create</a>(\*\*<a href="src/together/types/beta/jig/volume_create_params.py">params</a>) -> <a href="./src/together/types/beta/jig/volume.py">Volume</a></code>
+- <code title="get /deployments/storage/volumes/{id}">client.beta.jig.volumes.<a href="./src/together/resources/beta/jig/volumes.py">retrieve</a>(id) -> <a href="./src/together/types/beta/jig/volume.py">Volume</a></code>
+- <code title="patch /deployments/storage/volumes/{id}">client.beta.jig.volumes.<a href="./src/together/resources/beta/jig/volumes.py">update</a>(id, \*\*<a href="src/together/types/beta/jig/volume_update_params.py">params</a>) -> <a href="./src/together/types/beta/jig/volume.py">Volume</a></code>
+- <code title="get /deployments/storage/volumes">client.beta.jig.volumes.<a href="./src/together/resources/beta/jig/volumes.py">list</a>() -> <a href="./src/together/types/beta/jig/volume_list_response.py">VolumeListResponse</a></code>
+- <code title="delete /deployments/storage/volumes/{id}">client.beta.jig.volumes.<a href="./src/together/resources/beta/jig/volumes.py">delete</a>(id) -> object</code>
+
+### Secrets
+
+Types:
+
+```python
+from together.types.beta.jig import Secret, SecretListResponse
+```
+
+Methods:
+
+- <code title="post /deployments/secrets">client.beta.jig.secrets.<a href="./src/together/resources/beta/jig/secrets.py">create</a>(\*\*<a href="src/together/types/beta/jig/secret_create_params.py">params</a>) -> <a href="./src/together/types/beta/jig/secret.py">Secret</a></code>
+- <code title="get /deployments/secrets/{id}">client.beta.jig.secrets.<a href="./src/together/resources/beta/jig/secrets.py">retrieve</a>(id) -> <a href="./src/together/types/beta/jig/secret.py">Secret</a></code>
+- <code title="patch /deployments/secrets/{id}">client.beta.jig.secrets.<a href="./src/together/resources/beta/jig/secrets.py">update</a>(id, \*\*<a href="src/together/types/beta/jig/secret_update_params.py">params</a>) -> <a href="./src/together/types/beta/jig/secret.py">Secret</a></code>
+- <code title="get /deployments/secrets">client.beta.jig.secrets.<a href="./src/together/resources/beta/jig/secrets.py">list</a>() -> <a href="./src/together/types/beta/jig/secret_list_response.py">SecretListResponse</a></code>
+- <code title="delete /deployments/secrets/{id}">client.beta.jig.secrets.<a href="./src/together/resources/beta/jig/secrets.py">delete</a>(id) -> object</code>
+
 ## Clusters
 
 Types:
@@ -15,12 +84,12 @@ from together.types.beta import (
 
 Methods:
 
-- <code title="post /clusters">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">create</a>(\*\*<a href="src/together/types/beta/cluster_create_params.py">params</a>) -> <a href="./src/together/types/beta/cluster.py">Cluster</a></code>
-- <code title="get /clusters/{cluster_id}">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">retrieve</a>(cluster_id) -> <a href="./src/together/types/beta/cluster.py">Cluster</a></code>
-- <code title="put /clusters/{cluster_id}">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">update</a>(cluster_id, \*\*<a href="src/together/types/beta/cluster_update_params.py">params</a>) -> <a href="./src/together/types/beta/cluster.py">Cluster</a></code>
-- <code title="get /clusters">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">list</a>() -> <a href="./src/together/types/beta/cluster_list_response.py">ClusterListResponse</a></code>
-- <code title="delete /clusters/{cluster_id}">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">delete</a>(cluster_id) -> <a href="./src/together/types/beta/cluster_delete_response.py">ClusterDeleteResponse</a></code>
-- <code title="get /clusters/regions">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">list_regions</a>() -> <a href="./src/together/types/beta/cluster_list_regions_response.py">ClusterListRegionsResponse</a></code>
+- <code title="post /compute/clusters">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">create</a>(\*\*<a href="src/together/types/beta/cluster_create_params.py">params</a>) -> <a href="./src/together/types/beta/cluster.py">Cluster</a></code>
+- <code title="get /compute/clusters/{cluster_id}">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">retrieve</a>(cluster_id) -> <a href="./src/together/types/beta/cluster.py">Cluster</a></code>
+- <code title="put /compute/clusters/{cluster_id}">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">update</a>(cluster_id, \*\*<a href="src/together/types/beta/cluster_update_params.py">params</a>) -> <a href="./src/together/types/beta/cluster.py">Cluster</a></code>
+- <code title="get /compute/clusters">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">list</a>() -> <a href="./src/together/types/beta/cluster_list_response.py">ClusterListResponse</a></code>
+- <code title="delete /compute/clusters/{cluster_id}">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">delete</a>(cluster_id) -> <a href="./src/together/types/beta/cluster_delete_response.py">ClusterDeleteResponse</a></code>
+- <code title="get /compute/regions">client.beta.clusters.<a href="./src/together/resources/beta/clusters/clusters.py">list_regions</a>() -> <a href="./src/together/types/beta/cluster_list_regions_response.py">ClusterListRegionsResponse</a></code>
 
 ### Storage
 
@@ -32,11 +101,11 @@ from together.types.beta.clusters import ClusterStorage, StorageListResponse, St
 
 Methods:
 
-- <code title="post /clusters/storages">client.beta.clusters.storage.<a href="./src/together/resources/beta/clusters/storage.py">create</a>(\*\*<a href="src/together/types/beta/clusters/storage_create_params.py">params</a>) -> <a href="./src/together/types/beta/clusters/cluster_storage.py">ClusterStorage</a></code>
-- <code title="get /clusters/storages/{volume_id}">client.beta.clusters.storage.<a href="./src/together/resources/beta/clusters/storage.py">retrieve</a>(volume_id) -> <a href="./src/together/types/beta/clusters/cluster_storage.py">ClusterStorage</a></code>
-- <code title="put /clusters/storages">client.beta.clusters.storage.<a href="./src/together/resources/beta/clusters/storage.py">update</a>(\*\*<a href="src/together/types/beta/clusters/storage_update_params.py">params</a>) -> <a href="./src/together/types/beta/clusters/cluster_storage.py">ClusterStorage</a></code>
-- <code title="get /clusters/storages">client.beta.clusters.storage.<a href="./src/together/resources/beta/clusters/storage.py">list</a>() -> <a href="./src/together/types/beta/clusters/storage_list_response.py">StorageListResponse</a></code>
-- <code title="delete /clusters/storages/{volume_id}">client.beta.clusters.storage.<a href="./src/together/resources/beta/clusters/storage.py">delete</a>(volume_id) -> <a href="./src/together/types/beta/clusters/storage_delete_response.py">StorageDeleteResponse</a></code>
+- <code title="post /compute/clusters/storage/volumes">client.beta.clusters.storage.<a href="./src/together/resources/beta/clusters/storage.py">create</a>(\*\*<a href="src/together/types/beta/clusters/storage_create_params.py">params</a>) -> <a href="./src/together/types/beta/clusters/cluster_storage.py">ClusterStorage</a></code>
+- <code title="get /compute/clusters/storage/volumes/{volume_id}">client.beta.clusters.storage.<a href="./src/together/resources/beta/clusters/storage.py">retrieve</a>(volume_id) -> <a href="./src/together/types/beta/clusters/cluster_storage.py">ClusterStorage</a></code>
+- <code title="put /compute/clusters/storage/volumes">client.beta.clusters.storage.<a href="./src/together/resources/beta/clusters/storage.py">update</a>(\*\*<a href="src/together/types/beta/clusters/storage_update_params.py">params</a>) -> <a href="./src/together/types/beta/clusters/cluster_storage.py">ClusterStorage</a></code>
+- <code title="get /compute/clusters/storage/volumes">client.beta.clusters.storage.<a href="./src/together/resources/beta/clusters/storage.py">list</a>() -> <a href="./src/together/types/beta/clusters/storage_list_response.py">StorageListResponse</a></code>
+- <code title="delete /compute/clusters/storage/volumes/{volume_id}">client.beta.clusters.storage.<a href="./src/together/resources/beta/clusters/storage.py">delete</a>(volume_id) -> <a href="./src/together/types/beta/clusters/storage_delete_response.py">StorageDeleteResponse</a></code>
 
 # Chat
 
@@ -244,21 +313,20 @@ from together.types import ModelObject, ModelListResponse, ModelUploadResponse
 
 Methods:
 
-- <code title="get /models">client.models.<a href="./src/together/resources/models.py">list</a>(\*\*<a href="src/together/types/model_list_params.py">params</a>) -> <a href="./src/together/types/model_list_response.py">ModelListResponse</a></code>
-- <code title="post /models">client.models.<a href="./src/together/resources/models.py">upload</a>(\*\*<a href="src/together/types/model_upload_params.py">params</a>) -> <a href="./src/together/types/model_upload_response.py">ModelUploadResponse</a></code>
+- <code title="get /models">client.models.<a href="./src/together/resources/models/models.py">list</a>(\*\*<a href="src/together/types/model_list_params.py">params</a>) -> <a href="./src/together/types/model_list_response.py">ModelListResponse</a></code>
+- <code title="post /models">client.models.<a href="./src/together/resources/models/models.py">upload</a>(\*\*<a href="src/together/types/model_upload_params.py">params</a>) -> <a href="./src/together/types/model_upload_response.py">ModelUploadResponse</a></code>
 
-# Jobs
+## Uploads
 
 Types:
 
 ```python
-from together.types import JobRetrieveResponse, JobListResponse
+from together.types.models import UploadStatusResponse
 ```
 
 Methods:
 
-- <code title="get /jobs/{jobId}">client.jobs.<a href="./src/together/resources/jobs.py">retrieve</a>(job_id) -> <a href="./src/together/types/job_retrieve_response.py">JobRetrieveResponse</a></code>
-- <code title="get /jobs">client.jobs.<a href="./src/together/resources/jobs.py">list</a>() -> <a href="./src/together/types/job_list_response.py">JobListResponse</a></code>
+- <code title="get /jobs/{jobId}">client.models.uploads.<a href="./src/together/resources/models/uploads.py">status</a>(job_id) -> <a href="./src/together/types/models/upload_status_response.py">UploadStatusResponse</a></code>
 
 # Endpoints
 

{together-2.0.0a16 → together-2.0.0a18}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "together"
-version = "2.0.0-alpha.16"
+version = "2.0.0-alpha.18"
 description = "The official Python library for the together API"
 dynamic = ["readme"]
 license = "Apache-2.0"

{together-2.0.0a16 → together-2.0.0a18}/src/together/_base_client.py

@@ -86,6 +86,7 @@ from ._exceptions import (
     APIConnectionError,
     APIResponseValidationError,
 )
+from ._utils._json import openapi_dumps
 
 log: logging.Logger = logging.getLogger(__name__)
 
@@ -554,8 +555,10 @@ class BaseClient(Generic[_HttpxClientT, _DefaultStreamT]):
                 kwargs["content"] = options.content
             elif isinstance(json_data, bytes):
                 kwargs["content"] = json_data
-            else:
-                kwargs["json"] = json_data if is_given(json_data) else None
+            elif not files:
+                # Don't set content when JSON is sent as multipart/form-data,
+                # since httpx's content param overrides other body arguments
+                kwargs["content"] = openapi_dumps(json_data) if is_given(json_data) and json_data is not None else None
             kwargs["files"] = files
         else:
             headers.pop("Content-Type", None)

{together-2.0.0a16 → together-2.0.0a18}/src/together/_client.py

@@ -37,7 +37,6 @@ if TYPE_CHECKING:
     from .resources import (
         beta,
         chat,
-        jobs,
         audio,
         evals,
         files,
@@ -53,11 +52,9 @@ if TYPE_CHECKING:
         fine_tuning,
         code_interpreter,
     )
-    from .resources.jobs import JobsResource, AsyncJobsResource
     from .resources.evals import EvalsResource, AsyncEvalsResource
     from .resources.files import FilesResource, AsyncFilesResource
     from .resources.images import ImagesResource, AsyncImagesResource
-    from .resources.models import ModelsResource, AsyncModelsResource
     from .resources.rerank import RerankResource, AsyncRerankResource
     from .resources.videos import VideosResource, AsyncVideosResource
     from .resources.batches import BatchesResource, AsyncBatchesResource
@@ -69,6 +66,7 @@ if TYPE_CHECKING:
     from .resources.audio.audio import AudioResource, AsyncAudioResource
     from .resources.completions import CompletionsResource, AsyncCompletionsResource
     from .resources.fine_tuning import FineTuningResource, AsyncFineTuningResource
+    from .resources.models.models import ModelsResource, AsyncModelsResource
     from .resources.code_interpreter.code_interpreter import CodeInterpreterResource, AsyncCodeInterpreterResource
 
 __all__ = [
@@ -209,12 +207,6 @@ class Together(SyncAPIClient):
 
         return ModelsResource(self)
 
-    @cached_property
-    def jobs(self) -> JobsResource:
-        from .resources.jobs import JobsResource
-
-        return JobsResource(self)
-
     @cached_property
     def endpoints(self) -> EndpointsResource:
         from .resources.endpoints import EndpointsResource
@@ -486,12 +478,6 @@ class AsyncTogether(AsyncAPIClient):
 
         return AsyncModelsResource(self)
 
-    @cached_property
-    def jobs(self) -> AsyncJobsResource:
-        from .resources.jobs import AsyncJobsResource
-
-        return AsyncJobsResource(self)
-
     @cached_property
     def endpoints(self) -> AsyncEndpointsResource:
         from .resources.endpoints import AsyncEndpointsResource
@@ -709,12 +695,6 @@ class TogetherWithRawResponse:
 
         return ModelsResourceWithRawResponse(self._client.models)
 
-    @cached_property
-    def jobs(self) -> jobs.JobsResourceWithRawResponse:
-        from .resources.jobs import JobsResourceWithRawResponse
-
-        return JobsResourceWithRawResponse(self._client.jobs)
-
     @cached_property
     def endpoints(self) -> endpoints.EndpointsResourceWithRawResponse:
         from .resources.endpoints import EndpointsResourceWithRawResponse
@@ -818,12 +798,6 @@ class AsyncTogetherWithRawResponse:
 
         return AsyncModelsResourceWithRawResponse(self._client.models)
 
-    @cached_property
-    def jobs(self) -> jobs.AsyncJobsResourceWithRawResponse:
-        from .resources.jobs import AsyncJobsResourceWithRawResponse
-
-        return AsyncJobsResourceWithRawResponse(self._client.jobs)
-
     @cached_property
     def endpoints(self) -> endpoints.AsyncEndpointsResourceWithRawResponse:
         from .resources.endpoints import AsyncEndpointsResourceWithRawResponse
@@ -927,12 +901,6 @@ class TogetherWithStreamedResponse:
 
         return ModelsResourceWithStreamingResponse(self._client.models)
 
-    @cached_property
-    def jobs(self) -> jobs.JobsResourceWithStreamingResponse:
-        from .resources.jobs import JobsResourceWithStreamingResponse
-
-        return JobsResourceWithStreamingResponse(self._client.jobs)
-
     @cached_property
     def endpoints(self) -> endpoints.EndpointsResourceWithStreamingResponse:
         from .resources.endpoints import EndpointsResourceWithStreamingResponse
@@ -1036,12 +1004,6 @@ class AsyncTogetherWithStreamedResponse:
 
         return AsyncModelsResourceWithStreamingResponse(self._client.models)
 
-    @cached_property
-    def jobs(self) -> jobs.AsyncJobsResourceWithStreamingResponse:
-        from .resources.jobs import AsyncJobsResourceWithStreamingResponse
-
-        return AsyncJobsResourceWithStreamingResponse(self._client.jobs)
-
     @cached_property
     def endpoints(self) -> endpoints.AsyncEndpointsResourceWithStreamingResponse:
         from .resources.endpoints import AsyncEndpointsResourceWithStreamingResponse

{together-2.0.0a16 → together-2.0.0a18}/src/together/_compat.py

@@ -139,6 +139,7 @@ def model_dump(
     exclude_defaults: bool = False,
     warnings: bool = True,
     mode: Literal["json", "python"] = "python",
+    by_alias: bool | None = None,
 ) -> dict[str, Any]:
     if (not PYDANTIC_V1) or hasattr(model, "model_dump"):
         return model.model_dump(
@@ -148,13 +149,12 @@ def model_dump(
             exclude_defaults=exclude_defaults,
             # warnings are not supported in Pydantic v1
             warnings=True if PYDANTIC_V1 else warnings,
+            by_alias=by_alias,
         )
     return cast(
         "dict[str, Any]",
         model.dict(  # pyright: ignore[reportDeprecated, reportUnnecessaryCast]
-            exclude=exclude,
-            exclude_unset=exclude_unset,
-            exclude_defaults=exclude_defaults,
+            exclude=exclude, exclude_unset=exclude_unset, exclude_defaults=exclude_defaults, by_alias=bool(by_alias)
         ),
     )
 

together-2.0.0a18/src/together/_utils/_json.py

@@ -0,0 +1,35 @@
+import json
+from typing import Any
+from datetime import datetime
+from typing_extensions import override
+
+import pydantic
+
+from .._compat import model_dump
+
+
+def openapi_dumps(obj: Any) -> bytes:
+    """
+    Serialize an object to UTF-8 encoded JSON bytes.
+
+    Extends the standard json.dumps with support for additional types
+    commonly used in the SDK, such as `datetime`, `pydantic.BaseModel`, etc.
+    """
+    return json.dumps(
+        obj,
+        cls=_CustomEncoder,
+        # Uses the same defaults as httpx's JSON serialization
+        ensure_ascii=False,
+        separators=(",", ":"),
+        allow_nan=False,
+    ).encode()
+
+
+class _CustomEncoder(json.JSONEncoder):
+    @override
+    def default(self, o: Any) -> Any:
+        if isinstance(o, datetime):
+            return o.isoformat()
+        if isinstance(o, pydantic.BaseModel):
+            return model_dump(o, exclude_unset=True, mode="json", by_alias=True)
+        return super().default(o)

{together-2.0.0a16 → together-2.0.0a18}/src/together/_version.py

@@ -1,4 +1,4 @@
 # File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
 
 __title__ = "together"
-__version__ = "2.0.0-alpha.16"  # x-release-please-version
+__version__ = "2.0.0-alpha.18"  # x-release-please-version

{together-2.0.0a16 → together-2.0.0a18}/src/together/lib/cli/api/endpoints/create.py

@@ -3,7 +3,6 @@ from __future__ import annotations
 import sys
 
 import click
-from rich import print
 
 from together import APIError, Together, omit
 from together.lib.cli.api._utils import handle_api_errors
@@ -31,7 +30,7 @@ from .hardware import hardware
 )
 @click.option(
     "--gpu",
-    type=click.Choice(["h100", "a100", "l40", "l40s", "rtx-6000"]),
+    type=click.Choice(["b200", "h200", "h100", "a100", "l40", "l40s", "rtx-6000"]),
     required=True,
     help="GPU type to use for inference",
 )
@@ -48,7 +47,7 @@ from .hardware import hardware
 @click.option(
     "--no-prompt-cache",
     is_flag=True,
-    help="Disable the prompt cache for this endpoint",
+    help="Deprecated and no longer has any effect.",
 )
 @click.option(
     "--no-speculative-decoding",
@@ -95,6 +94,8 @@ def create(
     client: Together = ctx.obj
     # Map GPU types to their full hardware ID names
     gpu_map = {
+        "b200": "nvidia_b200_180gb_sxm",
+        "h200": "nvidia_h200_140gb_sxm",
         "h100": "nvidia_h100_80gb_sxm",
         "a100": "nvidia_a100_80gb_pcie" if gpu_count == 1 else "nvidia_a100_80gb_sxm",
         "l40": "nvidia_l40",
@@ -102,6 +103,9 @@ def create(
         "rtx-6000": "nvidia_rtx_6000_ada",
     }
 
+    if no_prompt_cache is not None:
+        click.echo("Warning: --no-prompt-cache is deprecated and no longer has any effect.", err=True)
+
     hardware_id = f"{gpu_count}x_{gpu_map[gpu]}"
 
     try:
@@ -113,15 +117,19 @@ def create(
                 "max_replicas": max_replicas,
             },
             display_name=display_name or omit,
-            disable_prompt_cache=no_prompt_cache or omit,
             disable_speculative_decoding=no_speculative_decoding or omit,
             state="STOPPED" if no_auto_start else "STARTED",
             inactive_timeout=inactive_timeout,
             extra_query={"availability_zone": availability_zone or omit},
         )
     except APIError as e:
-        if "check the hardware api" in str(e.args[0]).lower() or "invalid hardware provided" in str(e.args[0]).lower():
-            print("Invalid hardware provided")
+        if (
+            "check the hardware api" in str(e.args[0]).lower()
+            or "invalid hardware provided" in str(e.args[0]).lower()
+            or "the selected configuration" in str(e.args[0]).lower()
+        ):
+            click.secho("Invalid hardware selected.", fg="red", err=True)
+            click.echo("\nAvailable hardware options:")
             ctx.invoke(hardware, available=True, model=model, json=False)
             sys.exit(1)
         raise e
@@ -134,8 +142,6 @@ def create(
     click.echo(f"  Hardware: {hardware_id}", err=True)
     if display_name:
         click.echo(f"  Display name: {display_name}", err=True)
-    if no_prompt_cache:
-        click.echo("  Prompt cache: disabled", err=True)
     if no_speculative_decoding:
         click.echo("  Speculative decoding: disabled", err=True)
     if no_auto_start:

together-2.0.0a18/src/together/lib/cli/api/endpoints/hardware.py

@@ -0,0 +1,71 @@
+from __future__ import annotations
+
+import re
+import json as json_lib
+from typing import Any, Dict, List
+
+import click
+from tabulate import tabulate
+
+from together import Together, omit
+from together.types import HardwareListResponse
+from together.lib.cli.api._utils import handle_api_errors
+from together.lib.utils.serializer import datetime_serializer
+
+
+@click.command()
+@click.option("--model", help="Filter hardware options by model")
+@click.option("--json", is_flag=True, help="Print output in JSON format")
+@click.option(
+    "--available",
+    is_flag=True,
+    help="Print only available hardware options (can only be used if model is passed in)",
+)
+@click.pass_obj
+@handle_api_errors("Endpoints")
+def hardware(client: Together, model: str | None, json: bool, available: bool) -> None:
+    """List all available hardware options, optionally filtered by model."""
+    hardware_options = client.hardware.list(model=model or omit)
+
+    if available:
+        hardware_options.data = [
+            hardware
+            for hardware in hardware_options.data
+            if hardware.availability is not None and hardware.availability.status == "available"
+        ]
+    if json:
+        json_output = [hardware.model_dump() for hardware in hardware_options.data]
+        click.echo(json_lib.dumps(json_output, default=datetime_serializer, indent=2))
+    else:
+        _format_hardware_options(hardware_options, show_availability=model is not None)
+
+
+def _format_hardware_options(hardware_options: HardwareListResponse, show_availability: bool = True) -> None:
+    display_list: List[Dict[str, Any]] = []
+
+    for hw in hardware_options.data:
+        data = {
+            "Hardware ID": hw.id,
+            "GPU": re.sub(r"\-\d+[a-zA-Z][a-zA-Z]$", "", hw.specs.gpu_type)
+            if hw.specs and hw.specs.gpu_type
+            else "N/A",
+            "Memory": f"{int(hw.specs.gpu_memory)}GB" if hw.specs else "N/A",
+            "Count": hw.specs.gpu_count if hw.specs else "N/A",
+            "Price (per minute)": (f"${hw.pricing.cents_per_minute / 100:.2f}" if hw.pricing else "N/A"),
+        }
+
+        if show_availability:
+            status_display = "—"
+            if hw.availability:
+                status = hw.availability.status
+                # Add visual indicators for status
+                if status == "available":
+                    status_display = click.style("✓ available", fg="green")
+                elif status == "unavailable":
+                    status_display = click.style("✗ unavailable", fg="red")
+                else:  # insufficient
+                    status_display = click.style("⚠ insufficient", fg="yellow")
+                data["availability"] = status_display
+        display_list.append(data)
+
+    click.echo(tabulate(display_list, headers="keys", numalign="left"))