diff --git a/.release-please-manifest.json b/.release-please-manifest.json
index 54198710..9fc99f6a 100644
--- a/.release-please-manifest.json
+++ b/.release-please-manifest.json
@@ -1,3 +1,3 @@
 {
-  ".": "0.2.0-alpha.26"
+  ".": "0.2.0-alpha.31"
 }
\ No newline at end of file
diff --git a/.stats.yml b/.stats.yml
index 6a8c1428..dd473053 100644
--- a/.stats.yml
+++ b/.stats.yml
@@ -1 +1 @@
-configured_endpoints: 13
+configured_endpoints: 14
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index b47733a9..1a053ce9 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -2,9 +2,13 @@
 
 ### With Rye
 
-We use [Rye](https://rye.astral.sh/) to manage dependencies so we highly recommend [installing it](https://rye.astral.sh/guide/installation/) as it will automatically provision a Python environment with the expected Python version.
+We use [Rye](https://rye.astral.sh/) to manage dependencies because it will automatically provision a Python environment with the expected Python version. To set it up, run:
 
-After installing Rye, you'll just have to run this command:
+```sh
+$ ./scripts/bootstrap
+```
+
+Or [install Rye manually](https://rye.astral.sh/guide/installation/) and run:
 
 ```sh
 $ rye sync --all-features
@@ -31,25 +35,25 @@ $ pip install -r requirements-dev.lock
 
 ## Modifying/Adding code
 
-Most of the SDK is generated code, and any modified code will be overridden on the next generation. The
-`src/openlayer/lib/` and `examples/` directories are exceptions and will never be overridden.
+Most of the SDK is generated code. Modifications to code will be persisted between generations, but may
+result in merge conflicts between manual patches and changes from the generator. The generator will never
+modify the contents of the `src/openlayer/lib/` and `examples/` directories.
 
 ## Adding and running examples
 
-All files in the `examples/` directory are not modified by the Stainless generator and can be freely edited or
-added to.
+All files in the `examples/` directory are not modified by the generator and can be freely edited or added to.
 
-```bash
+```py
 # add an example to examples/<your-example>.py
 
 #!/usr/bin/env -S rye run python
 …
 ```
 
-```
-chmod +x examples/<your-example>.py
+```sh
+$ chmod +x examples/<your-example>.py
 # run the example against your api
-./examples/<your-example>.py
+$ ./examples/<your-example>.py
 ```
 
 ## Using the repository from source
@@ -58,8 +62,8 @@ If you’d like to use the repository from source, you can either install from g
 
 To install via git:
 
-```bash
-pip install git+ssh://git@github.com/openlayer-ai/openlayer-python.git
+```sh
+$ pip install git+ssh://git@github.com/openlayer-ai/openlayer-python.git
 ```
 
 Alternatively, you can build from source and install the wheel file:
@@ -68,29 +72,29 @@ Building this package will create two files in the `dist/` directory, a `.tar.gz
 
 To create a distributable version of the library, all you have to do is run this command:
 
-```bash
-rye build
+```sh
+$ rye build
 # or
-python -m build
+$ python -m build
 ```
 
 Then to install:
 
 ```sh
-pip install ./path-to-wheel-file.whl
+$ pip install ./path-to-wheel-file.whl
 ```
 
 ## Running tests
 
 Most tests require you to [set up a mock server](https://github.com/stoplightio/prism) against the OpenAPI spec to run the tests.
 
-```bash
+```sh
 # you will need npm installed
-npx prism mock path/to/your/openapi.yml
+$ npx prism mock path/to/your/openapi.yml
 ```
 
-```bash
-rye run pytest
+```sh
+$ ./scripts/test
 ```
 
 ## Linting and formatting
@@ -100,14 +104,14 @@ This repository uses [ruff](https://github.com/astral-sh/ruff) and
 
 To lint:
 
-```bash
-rye run lint
+```sh
+$ ./scripts/lint
 ```
 
 To format and fix all ruff issues automatically:
 
-```bash
-rye run format
+```sh
+$ ./scripts/format
 ```
 
 ## Publishing and releases
diff --git a/README.md b/README.md
index 1d316a13..93efe936 100644
--- a/README.md
+++ b/README.md
@@ -32,7 +32,7 @@ client = Openlayer(
     api_key=os.environ.get("OPENLAYER_API_KEY"),
 )
 
-data_stream_response = client.inference_pipelines.data.stream(
+response = client.inference_pipelines.data.stream(
     inference_pipeline_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
     config={
         "input_variable_names": ["user_query"],
@@ -47,11 +47,11 @@ data_stream_response = client.inference_pipelines.data.stream(
             "output": "42",
             "tokens": 7,
             "cost": 0.02,
-            "timestamp": 1620000000,
+            "timestamp": 1610000000,
         }
     ],
 )
-print(data_stream_response.success)
+print(response.success)
 ```
 
 While you can provide an `api_key` keyword argument,
@@ -75,7 +75,7 @@ client = AsyncOpenlayer(
 
 
 async def main() -> None:
-    data_stream_response = await client.inference_pipelines.data.stream(
+    response = await client.inference_pipelines.data.stream(
         inference_pipeline_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
         config={
             "input_variable_names": ["user_query"],
@@ -90,11 +90,11 @@ async def main() -> None:
                 "output": "42",
                 "tokens": 7,
                 "cost": 0.02,
-                "timestamp": 1620000000,
+                "timestamp": 1610000000,
             }
         ],
     )
-    print(data_stream_response.success)
+    print(response.success)
 
 
 asyncio.run(main())
@@ -142,7 +142,7 @@ try:
                 "output": "42",
                 "tokens": 7,
                 "cost": 0.02,
-                "timestamp": 1620000000,
+                "timestamp": 1610000000,
             }
         ],
     )
@@ -203,7 +203,7 @@ client.with_options(max_retries=5).inference_pipelines.data.stream(
             "output": "42",
             "tokens": 7,
             "cost": 0.02,
-            "timestamp": 1620000000,
+            "timestamp": 1610000000,
         }
     ],
 )
@@ -244,7 +244,7 @@ client.with_options(timeout=5.0).inference_pipelines.data.stream(
             "output": "42",
             "tokens": 7,
             "cost": 0.02,
-            "timestamp": 1620000000,
+            "timestamp": 1610000000,
         }
     ],
 )
@@ -300,7 +300,7 @@ response = client.inference_pipelines.data.with_raw_response.stream(
         "output": "42",
         "tokens": 7,
         "cost": 0.02,
-        "timestamp": 1620000000,
+        "timestamp": 1610000000,
     }],
 )
 print(response.headers.get('X-My-Header'))
@@ -335,7 +335,7 @@ with client.inference_pipelines.data.with_streaming_response.stream(
             "output": "42",
             "tokens": 7,
             "cost": 0.02,
-            "timestamp": 1620000000,
+            "timestamp": 1610000000,
         }
     ],
 ) as response:
@@ -425,6 +425,21 @@ We take backwards-compatibility seriously and work hard to ensure you can rely o
 
 We are keen for your feedback; please open an [issue](https://www.github.com/openlayer-ai/openlayer-python/issues) with questions, bugs, or suggestions.
 
+### Determining the installed version
+
+If you've upgraded to the latest version but aren't seeing any new features you were expecting then your python environment is likely still using an older version.
+
+You can determine the version that is being used at runtime with:
+
+```py
+import openlayer
+print(openlayer.__version__)
+```
+
 ## Requirements
 
 Python 3.7 or higher.
+
+## Contributing
+
+See [the contributing documentation](./CONTRIBUTING.md).
diff --git a/api.md b/api.md
index 82e9d940..24e491a6 100644
--- a/api.md
+++ b/api.md
@@ -38,6 +38,16 @@ Methods:
 
 # Commits
 
+Types:
+
+```python
+from openlayer.types import CommitCreateResponse
+```
+
+Methods:
+
+- <code title="post /projects/{projectId}/versions">client.commits.<a href="./src/openlayer/resources/commits/commits.py">create</a>(project_id, \*\*<a href="src/openlayer/types/commit_create_params.py">params</a>) -> <a href="./src/openlayer/types/commit_create_response.py">CommitCreateResponse</a></code>
+
 ## TestResults
 
 Types:
diff --git a/pyproject.toml b/pyproject.toml
index 8b7d1f5e..6b481e75 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,6 @@
 [project]
 name = "openlayer"
-version = "0.2.0-alpha.26"
+version = "0.2.0-alpha.31"
 description = "The official Python library for the openlayer API"
 dynamic = ["readme"]
 license = "Apache-2.0"
@@ -15,7 +15,6 @@ dependencies = [
     "distro>=1.7.0, <2",
     "sniffio",
     "cached-property; python_version < '3.8'",
-
 ]
 requires-python = ">= 3.7"
 classifiers = [
@@ -36,8 +35,6 @@ classifiers = [
   "License :: OSI Approved :: Apache Software License"
 ]
 
-
-
 [project.urls]
 Homepage = "https://github.com/openlayer-ai/openlayer-python"
 Repository = "https://github.com/openlayer-ai/openlayer-python"
@@ -59,7 +56,6 @@ dev-dependencies = [
     "dirty-equals>=0.6.0",
     "importlib-metadata>=6.7.0",
     "rich>=13.7.1",
-
 ]
 
 [tool.rye.scripts]
@@ -67,11 +63,11 @@ format = { chain = [
   "format:ruff",
   "format:docs",
   "fix:ruff",
+  # run formatting again to fix any inconsistencies when imports are stripped
+  "format:ruff",
 ]}
-"format:black" = "black ."
 "format:docs" = "python scripts/utils/ruffen-docs.py README.md api.md"
 "format:ruff" = "ruff format"
-"format:isort" = "isort ."
 
 "lint" = { chain = [
   "check:ruff",
@@ -129,10 +125,6 @@ path = "README.md"
 pattern = '\[(.+?)\]\(((?!https?://)\S+?)\)'
 replacement = '[\1](https://github.com/openlayer-ai/openlayer-python/tree/main/\g<2>)'
 
-[tool.black]
-line-length = 120
-target-version = ["py37"]
-
 [tool.pytest.ini_options]
 testpaths = ["tests"]
 addopts = "--tb=short"
diff --git a/requirements-dev.lock b/requirements-dev.lock
index 4a4bf4e2..276215ee 100644
--- a/requirements-dev.lock
+++ b/requirements-dev.lock
@@ -16,8 +16,6 @@ anyio==4.4.0
     # via openlayer
 argcomplete==3.1.2
     # via nox
-attrs==23.1.0
-    # via pytest
 certifi==2023.7.22
     # via httpcore
     # via httpx
@@ -28,8 +26,9 @@ distlib==0.3.7
     # via virtualenv
 distro==1.8.0
     # via openlayer
-exceptiongroup==1.1.3
+exceptiongroup==1.2.2
     # via anyio
+    # via pytest
 filelock==3.12.4
     # via virtualenv
 h11==0.14.0
@@ -49,7 +48,7 @@ markdown-it-py==3.0.0
     # via rich
 mdurl==0.1.2
     # via markdown-it-py
-mypy==1.10.1
+mypy==1.11.2
 mypy-extensions==1.0.0
     # via mypy
 nodeenv==1.8.0
@@ -60,27 +59,25 @@ packaging==23.2
     # via pytest
 platformdirs==3.11.0
     # via virtualenv
-pluggy==1.3.0
-    # via pytest
-py==1.11.0
+pluggy==1.5.0
     # via pytest
-pydantic==2.7.1
+pydantic==2.9.2
     # via openlayer
-pydantic-core==2.18.2
+pydantic-core==2.23.4
     # via pydantic
 pygments==2.18.0
     # via rich
-pyright==1.1.374
-pytest==7.1.1
+pyright==1.1.380
+pytest==8.3.3
     # via pytest-asyncio
-pytest-asyncio==0.21.1
+pytest-asyncio==0.24.0
 python-dateutil==2.8.2
     # via time-machine
 pytz==2023.3.post1
     # via dirty-equals
 respx==0.20.2
 rich==13.7.1
-ruff==0.5.6
+ruff==0.6.9
 setuptools==68.2.2
     # via nodeenv
 six==1.16.0
@@ -90,10 +87,10 @@ sniffio==1.3.0
     # via httpx
     # via openlayer
 time-machine==2.9.0
-tomli==2.0.1
+tomli==2.0.2
     # via mypy
     # via pytest
-typing-extensions==4.8.0
+typing-extensions==4.12.2
     # via anyio
     # via mypy
     # via openlayer
diff --git a/requirements.lock b/requirements.lock
index 150a0182..4ef2ed1d 100644
--- a/requirements.lock
+++ b/requirements.lock
@@ -19,7 +19,7 @@ certifi==2023.7.22
     # via httpx
 distro==1.8.0
     # via openlayer
-exceptiongroup==1.1.3
+exceptiongroup==1.2.2
     # via anyio
 h11==0.14.0
     # via httpcore
@@ -30,15 +30,15 @@ httpx==0.25.2
 idna==3.4
     # via anyio
     # via httpx
-pydantic==2.7.1
+pydantic==2.9.2
     # via openlayer
-pydantic-core==2.18.2
+pydantic-core==2.23.4
     # via pydantic
 sniffio==1.3.0
     # via anyio
     # via httpx
     # via openlayer
-typing-extensions==4.8.0
+typing-extensions==4.12.2
     # via anyio
     # via openlayer
     # via pydantic
diff --git a/src/openlayer/_base_client.py b/src/openlayer/_base_client.py
index c47242a6..f37cfc90 100644
--- a/src/openlayer/_base_client.py
+++ b/src/openlayer/_base_client.py
@@ -143,6 +143,12 @@ def __init__(
         self.url = url
         self.params = params
 
+    @override
+    def __repr__(self) -> str:
+        if self.url:
+            return f"{self.__class__.__name__}(url={self.url})"
+        return f"{self.__class__.__name__}(params={self.params})"
+
 
 class BasePage(GenericModel, Generic[_T]):
     """
@@ -400,14 +406,7 @@ def _make_status_error(
     ) -> _exceptions.APIStatusError:
         raise NotImplementedError()
 
-    def _remaining_retries(
-        self,
-        remaining_retries: Optional[int],
-        options: FinalRequestOptions,
-    ) -> int:
-        return remaining_retries if remaining_retries is not None else options.get_max_retries(self.max_retries)
-
-    def _build_headers(self, options: FinalRequestOptions) -> httpx.Headers:
+    def _build_headers(self, options: FinalRequestOptions, *, retries_taken: int = 0) -> httpx.Headers:
         custom_headers = options.headers or {}
         headers_dict = _merge_mappings(self.default_headers, custom_headers)
         self._validate_headers(headers_dict, custom_headers)
@@ -419,6 +418,11 @@ def _build_headers(self, options: FinalRequestOptions) -> httpx.Headers:
         if idempotency_header and options.method.lower() != "get" and idempotency_header not in headers:
             headers[idempotency_header] = options.idempotency_key or self._idempotency_key()
 
+        # Don't set the retry count header if it was already set or removed by the caller. We check
+        # `custom_headers`, which can contain `Omit()`, instead of `headers` to account for the removal case.
+        if "x-stainless-retry-count" not in (header.lower() for header in custom_headers):
+            headers["x-stainless-retry-count"] = str(retries_taken)
+
         return headers
 
     def _prepare_url(self, url: str) -> URL:
@@ -440,6 +444,8 @@ def _make_sse_decoder(self) -> SSEDecoder | SSEBytesDecoder:
     def _build_request(
         self,
         options: FinalRequestOptions,
+        *,
+        retries_taken: int = 0,
     ) -> httpx.Request:
         if log.isEnabledFor(logging.DEBUG):
             log.debug("Request options: %s", model_dump(options, exclude_unset=True))
@@ -455,7 +461,7 @@ def _build_request(
             else:
                 raise RuntimeError(f"Unexpected JSON data type, {type(json_data)}, cannot merge with `extra_body`")
 
-        headers = self._build_headers(options)
+        headers = self._build_headers(options, retries_taken=retries_taken)
         params = _merge_mappings(self.default_query, options.params)
         content_type = headers.get("Content-Type")
         files = options.files
@@ -489,12 +495,17 @@ def _build_request(
             if not files:
                 files = cast(HttpxRequestFiles, ForceMultipartDict())
 
+        prepared_url = self._prepare_url(options.url)
+        if "_" in prepared_url.host:
+            # work around https://github.com/encode/httpx/discussions/2880
+            kwargs["extensions"] = {"sni_hostname": prepared_url.host.replace("_", "-")}
+
         # TODO: report this error to httpx
         return self._client.build_request(  # pyright: ignore[reportUnknownMemberType]
             headers=headers,
             timeout=self.timeout if isinstance(options.timeout, NotGiven) else options.timeout,
             method=options.method,
-            url=self._prepare_url(options.url),
+            url=prepared_url,
             # the `Query` type that we use is incompatible with qs'
             # `Params` type as it needs to be typed as `Mapping[str, object]`
             # so that passing a `TypedDict` doesn't cause an error.
@@ -684,7 +695,8 @@ def _calculate_retry_timeout(
         if retry_after is not None and 0 < retry_after <= 60:
             return retry_after
 
-        nb_retries = max_retries - remaining_retries
+        # Also cap retry count to 1000 to avoid any potential overflows with `pow`
+        nb_retries = min(max_retries - remaining_retries, 1000)
 
         # Apply exponential backoff, but not more than the max.
         sleep_seconds = min(INITIAL_RETRY_DELAY * pow(2.0, nb_retries), MAX_RETRY_DELAY)
@@ -933,12 +945,17 @@ def request(
         stream: bool = False,
         stream_cls: type[_StreamT] | None = None,
     ) -> ResponseT | _StreamT:
+        if remaining_retries is not None:
+            retries_taken = options.get_max_retries(self.max_retries) - remaining_retries
+        else:
+            retries_taken = 0
+
         return self._request(
             cast_to=cast_to,
             options=options,
             stream=stream,
             stream_cls=stream_cls,
-            remaining_retries=remaining_retries,
+            retries_taken=retries_taken,
         )
 
     def _request(
@@ -946,7 +963,7 @@ def _request(
         *,
         cast_to: Type[ResponseT],
         options: FinalRequestOptions,
-        remaining_retries: int | None,
+        retries_taken: int,
         stream: bool,
         stream_cls: type[_StreamT] | None,
     ) -> ResponseT | _StreamT:
@@ -958,8 +975,8 @@ def _request(
         cast_to = self._maybe_override_cast_to(cast_to, options)
         options = self._prepare_options(options)
 
-        retries = self._remaining_retries(remaining_retries, options)
-        request = self._build_request(options)
+        remaining_retries = options.get_max_retries(self.max_retries) - retries_taken
+        request = self._build_request(options, retries_taken=retries_taken)
         self._prepare_request(request)
 
         kwargs: HttpxSendArgs = {}
@@ -977,11 +994,11 @@ def _request(
         except httpx.TimeoutException as err:
             log.debug("Encountered httpx.TimeoutException", exc_info=True)
 
-            if retries > 0:
+            if remaining_retries > 0:
                 return self._retry_request(
                     input_options,
                     cast_to,
-                    retries,
+                    retries_taken=retries_taken,
                     stream=stream,
                     stream_cls=stream_cls,
                     response_headers=None,
@@ -992,11 +1009,11 @@ def _request(
         except Exception as err:
             log.debug("Encountered Exception", exc_info=True)
 
-            if retries > 0:
+            if remaining_retries > 0:
                 return self._retry_request(
                     input_options,
                     cast_to,
-                    retries,
+                    retries_taken=retries_taken,
                     stream=stream,
                     stream_cls=stream_cls,
                     response_headers=None,
@@ -1019,13 +1036,13 @@ def _request(
         except httpx.HTTPStatusError as err:  # thrown on 4xx and 5xx status code
             log.debug("Encountered httpx.HTTPStatusError", exc_info=True)
 
-            if retries > 0 and self._should_retry(err.response):
+            if remaining_retries > 0 and self._should_retry(err.response):
                 err.response.close()
                 return self._retry_request(
                     input_options,
                     cast_to,
-                    retries,
-                    err.response.headers,
+                    retries_taken=retries_taken,
+                    response_headers=err.response.headers,
                     stream=stream,
                     stream_cls=stream_cls,
                 )
@@ -1044,26 +1061,26 @@ def _request(
             response=response,
             stream=stream,
             stream_cls=stream_cls,
-            retries_taken=options.get_max_retries(self.max_retries) - retries,
+            retries_taken=retries_taken,
         )
 
     def _retry_request(
         self,
         options: FinalRequestOptions,
         cast_to: Type[ResponseT],
-        remaining_retries: int,
-        response_headers: httpx.Headers | None,
         *,
+        retries_taken: int,
+        response_headers: httpx.Headers | None,
         stream: bool,
         stream_cls: type[_StreamT] | None,
     ) -> ResponseT | _StreamT:
-        remaining = remaining_retries - 1
-        if remaining == 1:
+        remaining_retries = options.get_max_retries(self.max_retries) - retries_taken
+        if remaining_retries == 1:
             log.debug("1 retry left")
         else:
-            log.debug("%i retries left", remaining)
+            log.debug("%i retries left", remaining_retries)
 
-        timeout = self._calculate_retry_timeout(remaining, options, response_headers)
+        timeout = self._calculate_retry_timeout(remaining_retries, options, response_headers)
         log.info("Retrying request to %s in %f seconds", options.url, timeout)
 
         # In a synchronous context we are blocking the entire thread. Up to the library user to run the client in a
@@ -1073,7 +1090,7 @@ def _retry_request(
         return self._request(
             options=options,
             cast_to=cast_to,
-            remaining_retries=remaining,
+            retries_taken=retries_taken + 1,
             stream=stream,
             stream_cls=stream_cls,
         )
@@ -1491,12 +1508,17 @@ async def request(
         stream_cls: type[_AsyncStreamT] | None = None,
         remaining_retries: Optional[int] = None,
     ) -> ResponseT | _AsyncStreamT:
+        if remaining_retries is not None:
+            retries_taken = options.get_max_retries(self.max_retries) - remaining_retries
+        else:
+            retries_taken = 0
+
         return await self._request(
             cast_to=cast_to,
             options=options,
             stream=stream,
             stream_cls=stream_cls,
-            remaining_retries=remaining_retries,
+            retries_taken=retries_taken,
         )
 
     async def _request(
@@ -1506,7 +1528,7 @@ async def _request(
         *,
         stream: bool,
         stream_cls: type[_AsyncStreamT] | None,
-        remaining_retries: int | None,
+        retries_taken: int,
     ) -> ResponseT | _AsyncStreamT:
         if self._platform is None:
             # `get_platform` can make blocking IO calls so we
@@ -1521,8 +1543,8 @@ async def _request(
         cast_to = self._maybe_override_cast_to(cast_to, options)
         options = await self._prepare_options(options)
 
-        retries = self._remaining_retries(remaining_retries, options)
-        request = self._build_request(options)
+        remaining_retries = options.get_max_retries(self.max_retries) - retries_taken
+        request = self._build_request(options, retries_taken=retries_taken)
         await self._prepare_request(request)
 
         kwargs: HttpxSendArgs = {}
@@ -1538,11 +1560,11 @@ async def _request(
         except httpx.TimeoutException as err:
             log.debug("Encountered httpx.TimeoutException", exc_info=True)
 
-            if retries > 0:
+            if remaining_retries > 0:
                 return await self._retry_request(
                     input_options,
                     cast_to,
-                    retries,
+                    retries_taken=retries_taken,
                     stream=stream,
                     stream_cls=stream_cls,
                     response_headers=None,
@@ -1553,11 +1575,11 @@ async def _request(
         except Exception as err:
             log.debug("Encountered Exception", exc_info=True)
 
-            if retries > 0:
+            if remaining_retries > 0:
                 return await self._retry_request(
                     input_options,
                     cast_to,
-                    retries,
+                    retries_taken=retries_taken,
                     stream=stream,
                     stream_cls=stream_cls,
                     response_headers=None,
@@ -1575,13 +1597,13 @@ async def _request(
         except httpx.HTTPStatusError as err:  # thrown on 4xx and 5xx status code
             log.debug("Encountered httpx.HTTPStatusError", exc_info=True)
 
-            if retries > 0 and self._should_retry(err.response):
+            if remaining_retries > 0 and self._should_retry(err.response):
                 await err.response.aclose()
                 return await self._retry_request(
                     input_options,
                     cast_to,
-                    retries,
-                    err.response.headers,
+                    retries_taken=retries_taken,
+                    response_headers=err.response.headers,
                     stream=stream,
                     stream_cls=stream_cls,
                 )
@@ -1600,26 +1622,26 @@ async def _request(
             response=response,
             stream=stream,
             stream_cls=stream_cls,
-            retries_taken=options.get_max_retries(self.max_retries) - retries,
+            retries_taken=retries_taken,
         )
 
     async def _retry_request(
         self,
         options: FinalRequestOptions,
         cast_to: Type[ResponseT],
-        remaining_retries: int,
-        response_headers: httpx.Headers | None,
         *,
+        retries_taken: int,
+        response_headers: httpx.Headers | None,
         stream: bool,
         stream_cls: type[_AsyncStreamT] | None,
     ) -> ResponseT | _AsyncStreamT:
-        remaining = remaining_retries - 1
-        if remaining == 1:
+        remaining_retries = options.get_max_retries(self.max_retries) - retries_taken
+        if remaining_retries == 1:
             log.debug("1 retry left")
         else:
-            log.debug("%i retries left", remaining)
+            log.debug("%i retries left", remaining_retries)
 
-        timeout = self._calculate_retry_timeout(remaining, options, response_headers)
+        timeout = self._calculate_retry_timeout(remaining_retries, options, response_headers)
         log.info("Retrying request to %s in %f seconds", options.url, timeout)
 
         await anyio.sleep(timeout)
@@ -1627,7 +1649,7 @@ async def _retry_request(
         return await self._request(
             options=options,
             cast_to=cast_to,
-            remaining_retries=remaining,
+            retries_taken=retries_taken + 1,
             stream=stream,
             stream_cls=stream_cls,
         )
diff --git a/src/openlayer/_compat.py b/src/openlayer/_compat.py
index 21fe6941..d89920d9 100644
--- a/src/openlayer/_compat.py
+++ b/src/openlayer/_compat.py
@@ -133,15 +133,17 @@ def model_json(model: pydantic.BaseModel, *, indent: int | None = None) -> str:
 def model_dump(
     model: pydantic.BaseModel,
     *,
-    exclude: IncEx = None,
+    exclude: IncEx | None = None,
     exclude_unset: bool = False,
     exclude_defaults: bool = False,
+    warnings: bool = True,
 ) -> dict[str, Any]:
     if PYDANTIC_V2:
         return model.model_dump(
             exclude=exclude,
             exclude_unset=exclude_unset,
             exclude_defaults=exclude_defaults,
+            warnings=warnings,
         )
     return cast(
         "dict[str, Any]",
diff --git a/src/openlayer/_models.py b/src/openlayer/_models.py
index d386eaa3..42551b76 100644
--- a/src/openlayer/_models.py
+++ b/src/openlayer/_models.py
@@ -176,7 +176,7 @@ def __str__(self) -> str:
     # Based on https://github.com/samuelcolvin/pydantic/issues/1168#issuecomment-817742836.
     @classmethod
     @override
-    def construct(
+    def construct(  # pyright: ignore[reportIncompatibleMethodOverride]
         cls: Type[ModelT],
         _fields_set: set[str] | None = None,
         **values: object,
@@ -248,8 +248,8 @@ def model_dump(
             self,
             *,
             mode: Literal["json", "python"] | str = "python",
-            include: IncEx = None,
-            exclude: IncEx = None,
+            include: IncEx | None = None,
+            exclude: IncEx | None = None,
             by_alias: bool = False,
             exclude_unset: bool = False,
             exclude_defaults: bool = False,
@@ -303,8 +303,8 @@ def model_dump_json(
             self,
             *,
             indent: int | None = None,
-            include: IncEx = None,
-            exclude: IncEx = None,
+            include: IncEx | None = None,
+            exclude: IncEx | None = None,
             by_alias: bool = False,
             exclude_unset: bool = False,
             exclude_defaults: bool = False,
diff --git a/src/openlayer/_response.py b/src/openlayer/_response.py
index 364e7503..7234cd68 100644
--- a/src/openlayer/_response.py
+++ b/src/openlayer/_response.py
@@ -192,6 +192,9 @@ def _parse(self, *, to: type[_T] | None = None) -> R | _T:
         if cast_to == float:
             return cast(R, float(response.text))
 
+        if cast_to == bool:
+            return cast(R, response.text.lower() == "true")
+
         origin = get_origin(cast_to) or cast_to
 
         if origin == APIResponse:
diff --git a/src/openlayer/_types.py b/src/openlayer/_types.py
index 3618c229..4135ae9e 100644
--- a/src/openlayer/_types.py
+++ b/src/openlayer/_types.py
@@ -16,7 +16,7 @@
     Optional,
     Sequence,
 )
-from typing_extensions import Literal, Protocol, TypeAlias, TypedDict, override, runtime_checkable
+from typing_extensions import Set, Literal, Protocol, TypeAlias, TypedDict, override, runtime_checkable
 
 import httpx
 import pydantic
@@ -193,7 +193,9 @@ def get(self, __key: str) -> str | None: ...
 
 # Note: copied from Pydantic
 # https://github.com/pydantic/pydantic/blob/32ea570bf96e84234d2992e1ddf40ab8a565925a/pydantic/main.py#L49
-IncEx: TypeAlias = "set[int] | set[str] | dict[int, Any] | dict[str, Any] | None"
+IncEx: TypeAlias = Union[
+    Set[int], Set[str], Mapping[int, Union["IncEx", Literal[True]]], Mapping[str, Union["IncEx", Literal[True]]]
+]
 
 PostParser = Callable[[Any], Any]
 
diff --git a/src/openlayer/_utils/_utils.py b/src/openlayer/_utils/_utils.py
index 2fc5a1c6..0bba17ca 100644
--- a/src/openlayer/_utils/_utils.py
+++ b/src/openlayer/_utils/_utils.py
@@ -363,12 +363,13 @@ def file_from_path(path: str) -> FileTypes:
 
 def get_required_header(headers: HeadersLike, header: str) -> str:
     lower_header = header.lower()
-    if isinstance(headers, Mapping):
-        for k, v in headers.items():
+    if is_mapping_t(headers):
+        # mypy doesn't understand the type narrowing here
+        for k, v in headers.items():  # type: ignore
             if k.lower() == lower_header and isinstance(v, str):
                 return v
 
-    """ to deal with the case where the header looks like Stainless-Event-Id """
+    # to deal with the case where the header looks like Stainless-Event-Id
     intercaps_header = re.sub(r"([^\w])(\w)", lambda pat: pat.group(1) + pat.group(2).upper(), header.capitalize())
 
     for normalized_header in [header, lower_header, header.upper(), intercaps_header]:
diff --git a/src/openlayer/_version.py b/src/openlayer/_version.py
index 81d0ee49..26025116 100644
--- a/src/openlayer/_version.py
+++ b/src/openlayer/_version.py
@@ -1,4 +1,4 @@
 # File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
 
 __title__ = "openlayer"
-__version__ = "0.2.0-alpha.26"  # x-release-please-version
+__version__ = "0.2.0-alpha.31"  # x-release-please-version
diff --git a/src/openlayer/resources/commits/commits.py b/src/openlayer/resources/commits/commits.py
index e9c62f89..774ae94f 100644
--- a/src/openlayer/resources/commits/commits.py
+++ b/src/openlayer/resources/commits/commits.py
@@ -2,8 +2,24 @@
 
 from __future__ import annotations
 
+from typing import Optional
+
+import httpx
+
+from ...types import commit_create_params
+from ..._types import NOT_GIVEN, Body, Query, Headers, NotGiven
+from ..._utils import (
+    maybe_transform,
+    async_maybe_transform,
+)
 from ..._compat import cached_property
 from ..._resource import SyncAPIResource, AsyncAPIResource
+from ..._response import (
+    to_raw_response_wrapper,
+    to_streamed_response_wrapper,
+    async_to_raw_response_wrapper,
+    async_to_streamed_response_wrapper,
+)
 from .test_results import (
     TestResultsResource,
     AsyncTestResultsResource,
@@ -12,6 +28,8 @@
     TestResultsResourceWithStreamingResponse,
     AsyncTestResultsResourceWithStreamingResponse,
 )
+from ..._base_client import make_request_options
+from ...types.commit_create_response import CommitCreateResponse
 
 __all__ = ["CommitsResource", "AsyncCommitsResource"]
 
@@ -23,12 +41,77 @@ def test_results(self) -> TestResultsResource:
 
     @cached_property
     def with_raw_response(self) -> CommitsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return CommitsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> CommitsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return CommitsResourceWithStreamingResponse(self)
 
+    def create(
+        self,
+        project_id: str,
+        *,
+        commit: commit_create_params.Commit,
+        storage_uri: str,
+        archived: Optional[bool] | NotGiven = NOT_GIVEN,
+        deployment_status: str | NotGiven = NOT_GIVEN,
+        # Use the following arguments if you need to pass additional parameters to the API that aren't available via kwargs.
+        # The extra values given here take precedence over values defined on the client or passed to this method.
+        extra_headers: Headers | None = None,
+        extra_query: Query | None = None,
+        extra_body: Body | None = None,
+        timeout: float | httpx.Timeout | None | NotGiven = NOT_GIVEN,
+    ) -> CommitCreateResponse:
+        """
+        Create a new commit (project version) in a project.
+
+        Args:
+          commit: The details of a commit (project version).
+
+          storage_uri: The storage URI where the commit bundle is stored.
+
+          archived: Whether the commit is archived.
+
+          deployment_status: The deployment status associated with the commit's model.
+
+          extra_headers: Send extra headers
+
+          extra_query: Add additional query parameters to the request
+
+          extra_body: Add additional JSON properties to the request
+
+          timeout: Override the client-level default timeout for this request, in seconds
+        """
+        if not project_id:
+            raise ValueError(f"Expected a non-empty value for `project_id` but received {project_id!r}")
+        return self._post(
+            f"/projects/{project_id}/versions",
+            body=maybe_transform(
+                {
+                    "commit": commit,
+                    "storage_uri": storage_uri,
+                    "archived": archived,
+                    "deployment_status": deployment_status,
+                },
+                commit_create_params.CommitCreateParams,
+            ),
+            options=make_request_options(
+                extra_headers=extra_headers, extra_query=extra_query, extra_body=extra_body, timeout=timeout
+            ),
+            cast_to=CommitCreateResponse,
+        )
+
 
 class AsyncCommitsResource(AsyncAPIResource):
     @cached_property
@@ -37,17 +120,86 @@ def test_results(self) -> AsyncTestResultsResource:
 
     @cached_property
     def with_raw_response(self) -> AsyncCommitsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return AsyncCommitsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> AsyncCommitsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return AsyncCommitsResourceWithStreamingResponse(self)
 
+    async def create(
+        self,
+        project_id: str,
+        *,
+        commit: commit_create_params.Commit,
+        storage_uri: str,
+        archived: Optional[bool] | NotGiven = NOT_GIVEN,
+        deployment_status: str | NotGiven = NOT_GIVEN,
+        # Use the following arguments if you need to pass additional parameters to the API that aren't available via kwargs.
+        # The extra values given here take precedence over values defined on the client or passed to this method.
+        extra_headers: Headers | None = None,
+        extra_query: Query | None = None,
+        extra_body: Body | None = None,
+        timeout: float | httpx.Timeout | None | NotGiven = NOT_GIVEN,
+    ) -> CommitCreateResponse:
+        """
+        Create a new commit (project version) in a project.
+
+        Args:
+          commit: The details of a commit (project version).
+
+          storage_uri: The storage URI where the commit bundle is stored.
+
+          archived: Whether the commit is archived.
+
+          deployment_status: The deployment status associated with the commit's model.
+
+          extra_headers: Send extra headers
+
+          extra_query: Add additional query parameters to the request
+
+          extra_body: Add additional JSON properties to the request
+
+          timeout: Override the client-level default timeout for this request, in seconds
+        """
+        if not project_id:
+            raise ValueError(f"Expected a non-empty value for `project_id` but received {project_id!r}")
+        return await self._post(
+            f"/projects/{project_id}/versions",
+            body=await async_maybe_transform(
+                {
+                    "commit": commit,
+                    "storage_uri": storage_uri,
+                    "archived": archived,
+                    "deployment_status": deployment_status,
+                },
+                commit_create_params.CommitCreateParams,
+            ),
+            options=make_request_options(
+                extra_headers=extra_headers, extra_query=extra_query, extra_body=extra_body, timeout=timeout
+            ),
+            cast_to=CommitCreateResponse,
+        )
+
 
 class CommitsResourceWithRawResponse:
     def __init__(self, commits: CommitsResource) -> None:
         self._commits = commits
 
+        self.create = to_raw_response_wrapper(
+            commits.create,
+        )
+
     @cached_property
     def test_results(self) -> TestResultsResourceWithRawResponse:
         return TestResultsResourceWithRawResponse(self._commits.test_results)
@@ -57,6 +209,10 @@ class AsyncCommitsResourceWithRawResponse:
     def __init__(self, commits: AsyncCommitsResource) -> None:
         self._commits = commits
 
+        self.create = async_to_raw_response_wrapper(
+            commits.create,
+        )
+
     @cached_property
     def test_results(self) -> AsyncTestResultsResourceWithRawResponse:
         return AsyncTestResultsResourceWithRawResponse(self._commits.test_results)
@@ -66,6 +222,10 @@ class CommitsResourceWithStreamingResponse:
     def __init__(self, commits: CommitsResource) -> None:
         self._commits = commits
 
+        self.create = to_streamed_response_wrapper(
+            commits.create,
+        )
+
     @cached_property
     def test_results(self) -> TestResultsResourceWithStreamingResponse:
         return TestResultsResourceWithStreamingResponse(self._commits.test_results)
@@ -75,6 +235,10 @@ class AsyncCommitsResourceWithStreamingResponse:
     def __init__(self, commits: AsyncCommitsResource) -> None:
         self._commits = commits
 
+        self.create = async_to_streamed_response_wrapper(
+            commits.create,
+        )
+
     @cached_property
     def test_results(self) -> AsyncTestResultsResourceWithStreamingResponse:
         return AsyncTestResultsResourceWithStreamingResponse(self._commits.test_results)
diff --git a/src/openlayer/resources/commits/test_results.py b/src/openlayer/resources/commits/test_results.py
index 3fcba2fa..0d37c7e0 100644
--- a/src/openlayer/resources/commits/test_results.py
+++ b/src/openlayer/resources/commits/test_results.py
@@ -31,10 +31,21 @@ class TestResultsResource(SyncAPIResource):
 
     @cached_property
     def with_raw_response(self) -> TestResultsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return TestResultsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> TestResultsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return TestResultsResourceWithStreamingResponse(self)
 
     def list(
@@ -104,10 +115,21 @@ def list(
 class AsyncTestResultsResource(AsyncAPIResource):
     @cached_property
     def with_raw_response(self) -> AsyncTestResultsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return AsyncTestResultsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> AsyncTestResultsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return AsyncTestResultsResourceWithStreamingResponse(self)
 
     async def list(
diff --git a/src/openlayer/resources/inference_pipelines/data.py b/src/openlayer/resources/inference_pipelines/data.py
index 9adb0910..f8b4b547 100644
--- a/src/openlayer/resources/inference_pipelines/data.py
+++ b/src/openlayer/resources/inference_pipelines/data.py
@@ -29,10 +29,21 @@
 class DataResource(SyncAPIResource):
     @cached_property
     def with_raw_response(self) -> DataResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return DataResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> DataResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return DataResourceWithStreamingResponse(self)
 
     def stream(
@@ -88,10 +99,21 @@ def stream(
 class AsyncDataResource(AsyncAPIResource):
     @cached_property
     def with_raw_response(self) -> AsyncDataResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return AsyncDataResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> AsyncDataResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return AsyncDataResourceWithStreamingResponse(self)
 
     async def stream(
diff --git a/src/openlayer/resources/inference_pipelines/inference_pipelines.py b/src/openlayer/resources/inference_pipelines/inference_pipelines.py
index f64b9dea..bc0f2fe5 100644
--- a/src/openlayer/resources/inference_pipelines/inference_pipelines.py
+++ b/src/openlayer/resources/inference_pipelines/inference_pipelines.py
@@ -66,10 +66,21 @@ def test_results(self) -> TestResultsResource:
 
     @cached_property
     def with_raw_response(self) -> InferencePipelinesResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return InferencePipelinesResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> InferencePipelinesResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return InferencePipelinesResourceWithStreamingResponse(self)
 
     def retrieve(
@@ -212,10 +223,21 @@ def test_results(self) -> AsyncTestResultsResource:
 
     @cached_property
     def with_raw_response(self) -> AsyncInferencePipelinesResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return AsyncInferencePipelinesResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> AsyncInferencePipelinesResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return AsyncInferencePipelinesResourceWithStreamingResponse(self)
 
     async def retrieve(
diff --git a/src/openlayer/resources/inference_pipelines/rows.py b/src/openlayer/resources/inference_pipelines/rows.py
index d3407927..f763b1ab 100644
--- a/src/openlayer/resources/inference_pipelines/rows.py
+++ b/src/openlayer/resources/inference_pipelines/rows.py
@@ -29,10 +29,21 @@
 class RowsResource(SyncAPIResource):
     @cached_property
     def with_raw_response(self) -> RowsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return RowsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> RowsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return RowsResourceWithStreamingResponse(self)
 
     def update(
@@ -90,10 +101,21 @@ def update(
 class AsyncRowsResource(AsyncAPIResource):
     @cached_property
     def with_raw_response(self) -> AsyncRowsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return AsyncRowsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> AsyncRowsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return AsyncRowsResourceWithStreamingResponse(self)
 
     async def update(
diff --git a/src/openlayer/resources/inference_pipelines/test_results.py b/src/openlayer/resources/inference_pipelines/test_results.py
index 37d1fb8e..4bcb435e 100644
--- a/src/openlayer/resources/inference_pipelines/test_results.py
+++ b/src/openlayer/resources/inference_pipelines/test_results.py
@@ -31,10 +31,21 @@ class TestResultsResource(SyncAPIResource):
 
     @cached_property
     def with_raw_response(self) -> TestResultsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return TestResultsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> TestResultsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return TestResultsResourceWithStreamingResponse(self)
 
     def list(
@@ -102,10 +113,21 @@ def list(
 class AsyncTestResultsResource(AsyncAPIResource):
     @cached_property
     def with_raw_response(self) -> AsyncTestResultsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return AsyncTestResultsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> AsyncTestResultsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return AsyncTestResultsResourceWithStreamingResponse(self)
 
     async def list(
diff --git a/src/openlayer/resources/projects/commits.py b/src/openlayer/resources/projects/commits.py
index f6666180..fd16de8f 100644
--- a/src/openlayer/resources/projects/commits.py
+++ b/src/openlayer/resources/projects/commits.py
@@ -27,10 +27,21 @@
 class CommitsResource(SyncAPIResource):
     @cached_property
     def with_raw_response(self) -> CommitsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return CommitsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> CommitsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return CommitsResourceWithStreamingResponse(self)
 
     def list(
@@ -86,10 +97,21 @@ def list(
 class AsyncCommitsResource(AsyncAPIResource):
     @cached_property
     def with_raw_response(self) -> AsyncCommitsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return AsyncCommitsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> AsyncCommitsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return AsyncCommitsResourceWithStreamingResponse(self)
 
     async def list(
diff --git a/src/openlayer/resources/projects/inference_pipelines.py b/src/openlayer/resources/projects/inference_pipelines.py
index 6c8fff28..e8999bdf 100644
--- a/src/openlayer/resources/projects/inference_pipelines.py
+++ b/src/openlayer/resources/projects/inference_pipelines.py
@@ -30,10 +30,21 @@
 class InferencePipelinesResource(SyncAPIResource):
     @cached_property
     def with_raw_response(self) -> InferencePipelinesResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return InferencePipelinesResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> InferencePipelinesResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return InferencePipelinesResourceWithStreamingResponse(self)
 
     def create(
@@ -139,10 +150,21 @@ def list(
 class AsyncInferencePipelinesResource(AsyncAPIResource):
     @cached_property
     def with_raw_response(self) -> AsyncInferencePipelinesResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return AsyncInferencePipelinesResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> AsyncInferencePipelinesResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return AsyncInferencePipelinesResourceWithStreamingResponse(self)
 
     async def create(
diff --git a/src/openlayer/resources/projects/projects.py b/src/openlayer/resources/projects/projects.py
index fad7171a..e5e90392 100644
--- a/src/openlayer/resources/projects/projects.py
+++ b/src/openlayer/resources/projects/projects.py
@@ -55,10 +55,21 @@ def inference_pipelines(self) -> InferencePipelinesResource:
 
     @cached_property
     def with_raw_response(self) -> ProjectsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return ProjectsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> ProjectsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return ProjectsResourceWithStreamingResponse(self)
 
     def create(
@@ -175,10 +186,21 @@ def inference_pipelines(self) -> AsyncInferencePipelinesResource:
 
     @cached_property
     def with_raw_response(self) -> AsyncProjectsResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return AsyncProjectsResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> AsyncProjectsResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return AsyncProjectsResourceWithStreamingResponse(self)
 
     async def create(
diff --git a/src/openlayer/resources/storage/presigned_url.py b/src/openlayer/resources/storage/presigned_url.py
index ad2990e5..5fb6fa1c 100644
--- a/src/openlayer/resources/storage/presigned_url.py
+++ b/src/openlayer/resources/storage/presigned_url.py
@@ -27,10 +27,21 @@
 class PresignedURLResource(SyncAPIResource):
     @cached_property
     def with_raw_response(self) -> PresignedURLResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return PresignedURLResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> PresignedURLResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return PresignedURLResourceWithStreamingResponse(self)
 
     def create(
@@ -76,10 +87,21 @@ def create(
 class AsyncPresignedURLResource(AsyncAPIResource):
     @cached_property
     def with_raw_response(self) -> AsyncPresignedURLResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return AsyncPresignedURLResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> AsyncPresignedURLResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return AsyncPresignedURLResourceWithStreamingResponse(self)
 
     async def create(
diff --git a/src/openlayer/resources/storage/storage.py b/src/openlayer/resources/storage/storage.py
index 935bdc43..ea2a3c99 100644
--- a/src/openlayer/resources/storage/storage.py
+++ b/src/openlayer/resources/storage/storage.py
@@ -23,10 +23,21 @@ def presigned_url(self) -> PresignedURLResource:
 
     @cached_property
     def with_raw_response(self) -> StorageResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return StorageResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> StorageResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return StorageResourceWithStreamingResponse(self)
 
 
@@ -37,10 +48,21 @@ def presigned_url(self) -> AsyncPresignedURLResource:
 
     @cached_property
     def with_raw_response(self) -> AsyncStorageResourceWithRawResponse:
+        """
+        This property can be used as a prefix for any HTTP method call to return the
+        the raw response object instead of the parsed content.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#accessing-raw-response-data-eg-headers
+        """
         return AsyncStorageResourceWithRawResponse(self)
 
     @cached_property
     def with_streaming_response(self) -> AsyncStorageResourceWithStreamingResponse:
+        """
+        An alternative to `.with_raw_response` that doesn't eagerly read the response body.
+
+        For more information, see https://www.github.com/openlayer-ai/openlayer-python#with_streaming_response
+        """
         return AsyncStorageResourceWithStreamingResponse(self)
 
 
diff --git a/src/openlayer/types/__init__.py b/src/openlayer/types/__init__.py
index 58883aff..48381166 100644
--- a/src/openlayer/types/__init__.py
+++ b/src/openlayer/types/__init__.py
@@ -3,8 +3,10 @@
 from __future__ import annotations
 
 from .project_list_params import ProjectListParams as ProjectListParams
+from .commit_create_params import CommitCreateParams as CommitCreateParams
 from .project_create_params import ProjectCreateParams as ProjectCreateParams
 from .project_list_response import ProjectListResponse as ProjectListResponse
+from .commit_create_response import CommitCreateResponse as CommitCreateResponse
 from .project_create_response import ProjectCreateResponse as ProjectCreateResponse
 from .inference_pipeline_update_params import InferencePipelineUpdateParams as InferencePipelineUpdateParams
 from .inference_pipeline_update_response import InferencePipelineUpdateResponse as InferencePipelineUpdateResponse
diff --git a/src/openlayer/types/commit_create_params.py b/src/openlayer/types/commit_create_params.py
new file mode 100644
index 00000000..2a7d54de
--- /dev/null
+++ b/src/openlayer/types/commit_create_params.py
@@ -0,0 +1,29 @@
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from __future__ import annotations
+
+from typing import Optional
+from typing_extensions import Required, Annotated, TypedDict
+
+from .._utils import PropertyInfo
+
+__all__ = ["CommitCreateParams", "Commit"]
+
+
+class CommitCreateParams(TypedDict, total=False):
+    commit: Required[Commit]
+    """The details of a commit (project version)."""
+
+    storage_uri: Required[Annotated[str, PropertyInfo(alias="storageUri")]]
+    """The storage URI where the commit bundle is stored."""
+
+    archived: Optional[bool]
+    """Whether the commit is archived."""
+
+    deployment_status: Annotated[str, PropertyInfo(alias="deploymentStatus")]
+    """The deployment status associated with the commit's model."""
+
+
+class Commit(TypedDict, total=False):
+    message: Required[str]
+    """The commit message."""
diff --git a/src/openlayer/types/commit_create_response.py b/src/openlayer/types/commit_create_response.py
new file mode 100644
index 00000000..82bf6d16
--- /dev/null
+++ b/src/openlayer/types/commit_create_response.py
@@ -0,0 +1,106 @@
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from typing import Optional
+from datetime import datetime
+from typing_extensions import Literal
+
+from pydantic import Field as FieldInfo
+
+from .._models import BaseModel
+
+__all__ = ["CommitCreateResponse", "Commit", "Links"]
+
+
+class Commit(BaseModel):
+    id: str
+    """The commit id."""
+
+    author_id: str = FieldInfo(alias="authorId")
+    """The author id of the commit."""
+
+    file_size: Optional[int] = FieldInfo(alias="fileSize", default=None)
+    """The size of the commit bundle in bytes."""
+
+    message: str
+    """The commit message."""
+
+    ml_model_id: Optional[str] = FieldInfo(alias="mlModelId", default=None)
+    """The model id."""
+
+    storage_uri: str = FieldInfo(alias="storageUri")
+    """The storage URI where the commit bundle is stored."""
+
+    training_dataset_id: Optional[str] = FieldInfo(alias="trainingDatasetId", default=None)
+    """The training dataset id."""
+
+    validation_dataset_id: Optional[str] = FieldInfo(alias="validationDatasetId", default=None)
+    """The validation dataset id."""
+
+    date_created: Optional[datetime] = FieldInfo(alias="dateCreated", default=None)
+    """The commit creation date."""
+
+    git_commit_ref: Optional[str] = FieldInfo(alias="gitCommitRef", default=None)
+    """The ref of the corresponding git commit."""
+
+    git_commit_sha: Optional[int] = FieldInfo(alias="gitCommitSha", default=None)
+    """The SHA of the corresponding git commit."""
+
+    git_commit_url: Optional[str] = FieldInfo(alias="gitCommitUrl", default=None)
+    """The URL of the corresponding git commit."""
+
+
+class Links(BaseModel):
+    app: str
+
+
+class CommitCreateResponse(BaseModel):
+    id: str
+    """The project version (commit) id."""
+
+    commit: Commit
+    """The details of a commit (project version)."""
+
+    date_archived: Optional[datetime] = FieldInfo(alias="dateArchived", default=None)
+    """The commit archive date."""
+
+    date_created: datetime = FieldInfo(alias="dateCreated")
+    """The project version (commit) creation date."""
+
+    failing_goal_count: int = FieldInfo(alias="failingGoalCount")
+    """The number of tests that are failing for the commit."""
+
+    ml_model_id: Optional[str] = FieldInfo(alias="mlModelId", default=None)
+    """The model id."""
+
+    passing_goal_count: int = FieldInfo(alias="passingGoalCount")
+    """The number of tests that are passing for the commit."""
+
+    project_id: str = FieldInfo(alias="projectId")
+    """The project id."""
+
+    status: Literal["queued", "running", "paused", "failed", "completed", "unknown"]
+    """The commit status.
+
+    Initially, the commit is `queued`, then, it switches to `running`. Finally, it
+    can be `paused`, `failed`, or `completed`.
+    """
+
+    status_message: Optional[str] = FieldInfo(alias="statusMessage", default=None)
+    """The commit status message."""
+
+    total_goal_count: int = FieldInfo(alias="totalGoalCount")
+    """The total number of tests for the commit."""
+
+    training_dataset_id: Optional[str] = FieldInfo(alias="trainingDatasetId", default=None)
+    """The training dataset id."""
+
+    validation_dataset_id: Optional[str] = FieldInfo(alias="validationDatasetId", default=None)
+    """The validation dataset id."""
+
+    archived: Optional[bool] = None
+    """Whether the commit is archived."""
+
+    deployment_status: Optional[str] = FieldInfo(alias="deploymentStatus", default=None)
+    """The deployment status associated with the commit's model."""
+
+    links: Optional[Links] = None
diff --git a/src/openlayer/types/storage/presigned_url_create_response.py b/src/openlayer/types/storage/presigned_url_create_response.py
index 71791bbf..db578318 100644
--- a/src/openlayer/types/storage/presigned_url_create_response.py
+++ b/src/openlayer/types/storage/presigned_url_create_response.py
@@ -17,4 +17,4 @@ class PresignedURLCreateResponse(BaseModel):
     """The presigned url."""
 
     fields: Optional[object] = None
-    """Fields to include in the body of the upload. Only needed by s3."""
+    """Fields to include in the body of the upload. Only needed by s3"""
diff --git a/tests/api_resources/test_commits.py b/tests/api_resources/test_commits.py
new file mode 100644
index 00000000..15e0f5d9
--- /dev/null
+++ b/tests/api_resources/test_commits.py
@@ -0,0 +1,136 @@
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from __future__ import annotations
+
+import os
+from typing import Any, cast
+
+import pytest
+
+from openlayer import Openlayer, AsyncOpenlayer
+from tests.utils import assert_matches_type
+from openlayer.types import CommitCreateResponse
+
+base_url = os.environ.get("TEST_API_BASE_URL", "http://127.0.0.1:4010")
+
+
+class TestCommits:
+    parametrize = pytest.mark.parametrize("client", [False, True], indirect=True, ids=["loose", "strict"])
+
+    @parametrize
+    def test_method_create(self, client: Openlayer) -> None:
+        commit = client.commits.create(
+            project_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            commit={"message": "Updated the prompt."},
+            storage_uri="s3://...",
+        )
+        assert_matches_type(CommitCreateResponse, commit, path=["response"])
+
+    @parametrize
+    def test_method_create_with_all_params(self, client: Openlayer) -> None:
+        commit = client.commits.create(
+            project_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            commit={"message": "Updated the prompt."},
+            storage_uri="s3://...",
+            archived=False,
+            deployment_status="Deployed",
+        )
+        assert_matches_type(CommitCreateResponse, commit, path=["response"])
+
+    @parametrize
+    def test_raw_response_create(self, client: Openlayer) -> None:
+        response = client.commits.with_raw_response.create(
+            project_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            commit={"message": "Updated the prompt."},
+            storage_uri="s3://...",
+        )
+
+        assert response.is_closed is True
+        assert response.http_request.headers.get("X-Stainless-Lang") == "python"
+        commit = response.parse()
+        assert_matches_type(CommitCreateResponse, commit, path=["response"])
+
+    @parametrize
+    def test_streaming_response_create(self, client: Openlayer) -> None:
+        with client.commits.with_streaming_response.create(
+            project_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            commit={"message": "Updated the prompt."},
+            storage_uri="s3://...",
+        ) as response:
+            assert not response.is_closed
+            assert response.http_request.headers.get("X-Stainless-Lang") == "python"
+
+            commit = response.parse()
+            assert_matches_type(CommitCreateResponse, commit, path=["response"])
+
+        assert cast(Any, response.is_closed) is True
+
+    @parametrize
+    def test_path_params_create(self, client: Openlayer) -> None:
+        with pytest.raises(ValueError, match=r"Expected a non-empty value for `project_id` but received ''"):
+            client.commits.with_raw_response.create(
+                project_id="",
+                commit={"message": "Updated the prompt."},
+                storage_uri="s3://...",
+            )
+
+
+class TestAsyncCommits:
+    parametrize = pytest.mark.parametrize("async_client", [False, True], indirect=True, ids=["loose", "strict"])
+
+    @parametrize
+    async def test_method_create(self, async_client: AsyncOpenlayer) -> None:
+        commit = await async_client.commits.create(
+            project_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            commit={"message": "Updated the prompt."},
+            storage_uri="s3://...",
+        )
+        assert_matches_type(CommitCreateResponse, commit, path=["response"])
+
+    @parametrize
+    async def test_method_create_with_all_params(self, async_client: AsyncOpenlayer) -> None:
+        commit = await async_client.commits.create(
+            project_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            commit={"message": "Updated the prompt."},
+            storage_uri="s3://...",
+            archived=False,
+            deployment_status="Deployed",
+        )
+        assert_matches_type(CommitCreateResponse, commit, path=["response"])
+
+    @parametrize
+    async def test_raw_response_create(self, async_client: AsyncOpenlayer) -> None:
+        response = await async_client.commits.with_raw_response.create(
+            project_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            commit={"message": "Updated the prompt."},
+            storage_uri="s3://...",
+        )
+
+        assert response.is_closed is True
+        assert response.http_request.headers.get("X-Stainless-Lang") == "python"
+        commit = await response.parse()
+        assert_matches_type(CommitCreateResponse, commit, path=["response"])
+
+    @parametrize
+    async def test_streaming_response_create(self, async_client: AsyncOpenlayer) -> None:
+        async with async_client.commits.with_streaming_response.create(
+            project_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            commit={"message": "Updated the prompt."},
+            storage_uri="s3://...",
+        ) as response:
+            assert not response.is_closed
+            assert response.http_request.headers.get("X-Stainless-Lang") == "python"
+
+            commit = await response.parse()
+            assert_matches_type(CommitCreateResponse, commit, path=["response"])
+
+        assert cast(Any, response.is_closed) is True
+
+    @parametrize
+    async def test_path_params_create(self, async_client: AsyncOpenlayer) -> None:
+        with pytest.raises(ValueError, match=r"Expected a non-empty value for `project_id` but received ''"):
+            await async_client.commits.with_raw_response.create(
+                project_id="",
+                commit={"message": "Updated the prompt."},
+                storage_uri="s3://...",
+            )
diff --git a/tests/conftest.py b/tests/conftest.py
index 0857c182..554ab710 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -1,11 +1,11 @@
 from __future__ import annotations
 
 import os
-import asyncio
 import logging
 from typing import TYPE_CHECKING, Iterator, AsyncIterator
 
 import pytest
+from pytest_asyncio import is_async_test
 
 from openlayer import Openlayer, AsyncOpenlayer
 
@@ -17,11 +17,13 @@
 logging.getLogger("openlayer").setLevel(logging.DEBUG)
 
 
-@pytest.fixture(scope="session")
-def event_loop() -> Iterator[asyncio.AbstractEventLoop]:
-    loop = asyncio.new_event_loop()
-    yield loop
-    loop.close()
+# automatically add `pytest.mark.asyncio()` to all of our async tests
+# so we don't have to add that boilerplate everywhere
+def pytest_collection_modifyitems(items: list[pytest.Function]) -> None:
+    pytest_asyncio_tests = (item for item in items if is_async_test(item))
+    session_scope_marker = pytest.mark.asyncio(loop_scope="session")
+    for async_test in pytest_asyncio_tests:
+        async_test.add_marker(session_scope_marker, append=False)
 
 
 base_url = os.environ.get("TEST_API_BASE_URL", "http://127.0.0.1:4010")
diff --git a/tests/test_client.py b/tests/test_client.py
index 7b312411..b57e50db 100644
--- a/tests/test_client.py
+++ b/tests/test_client.py
@@ -10,6 +10,7 @@
 import tracemalloc
 from typing import Any, Union, cast
 from unittest import mock
+from typing_extensions import Literal
 
 import httpx
 import pytest
@@ -701,6 +702,7 @@ class Model(BaseModel):
             [3, "", 0.5],
             [2, "", 0.5 * 2.0],
             [1, "", 0.5 * 4.0],
+            [-1100, "", 7.8],  # test large number potentially overflowing
         ],
     )
     @mock.patch("time.time", mock.MagicMock(return_value=1696004797))
@@ -738,7 +740,7 @@ def test_retrying_timeout_errors_doesnt_leak(self, respx_mock: MockRouter) -> No
                                 "output": "42",
                                 "tokens": 7,
                                 "cost": 0.02,
-                                "timestamp": 1620000000,
+                                "timestamp": 1610000000,
                             }
                         ],
                     ),
@@ -775,7 +777,7 @@ def test_retrying_status_errors_doesnt_leak(self, respx_mock: MockRouter) -> Non
                                 "output": "42",
                                 "tokens": 7,
                                 "cost": 0.02,
-                                "timestamp": 1620000000,
+                                "timestamp": 1610000000,
                             }
                         ],
                     ),
@@ -789,7 +791,14 @@ def test_retrying_status_errors_doesnt_leak(self, respx_mock: MockRouter) -> Non
     @pytest.mark.parametrize("failures_before_success", [0, 2, 4])
     @mock.patch("openlayer._base_client.BaseClient._calculate_retry_timeout", _low_retry_timeout)
     @pytest.mark.respx(base_url=base_url)
-    def test_retries_taken(self, client: Openlayer, failures_before_success: int, respx_mock: MockRouter) -> None:
+    @pytest.mark.parametrize("failure_mode", ["status", "exception"])
+    def test_retries_taken(
+        self,
+        client: Openlayer,
+        failures_before_success: int,
+        failure_mode: Literal["status", "exception"],
+        respx_mock: MockRouter,
+    ) -> None:
         client = client.with_options(max_retries=4)
 
         nb_retries = 0
@@ -798,6 +807,8 @@ def retry_handler(_request: httpx.Request) -> httpx.Response:
             nonlocal nb_retries
             if nb_retries < failures_before_success:
                 nb_retries += 1
+                if failure_mode == "exception":
+                    raise RuntimeError("oops")
                 return httpx.Response(500)
             return httpx.Response(200)
 
@@ -820,6 +831,83 @@ def retry_handler(_request: httpx.Request) -> httpx.Response:
         )
 
         assert response.retries_taken == failures_before_success
+        assert int(response.http_request.headers.get("x-stainless-retry-count")) == failures_before_success
+
+    @pytest.mark.parametrize("failures_before_success", [0, 2, 4])
+    @mock.patch("openlayer._base_client.BaseClient._calculate_retry_timeout", _low_retry_timeout)
+    @pytest.mark.respx(base_url=base_url)
+    def test_omit_retry_count_header(
+        self, client: Openlayer, failures_before_success: int, respx_mock: MockRouter
+    ) -> None:
+        client = client.with_options(max_retries=4)
+
+        nb_retries = 0
+
+        def retry_handler(_request: httpx.Request) -> httpx.Response:
+            nonlocal nb_retries
+            if nb_retries < failures_before_success:
+                nb_retries += 1
+                return httpx.Response(500)
+            return httpx.Response(200)
+
+        respx_mock.post("/inference-pipelines/182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e/data-stream").mock(
+            side_effect=retry_handler
+        )
+
+        response = client.inference_pipelines.data.with_raw_response.stream(
+            inference_pipeline_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            config={"output_column_name": "output"},
+            rows=[
+                {
+                    "user_query": "bar",
+                    "output": "bar",
+                    "tokens": "bar",
+                    "cost": "bar",
+                    "timestamp": "bar",
+                }
+            ],
+            extra_headers={"x-stainless-retry-count": Omit()},
+        )
+
+        assert len(response.http_request.headers.get_list("x-stainless-retry-count")) == 0
+
+    @pytest.mark.parametrize("failures_before_success", [0, 2, 4])
+    @mock.patch("openlayer._base_client.BaseClient._calculate_retry_timeout", _low_retry_timeout)
+    @pytest.mark.respx(base_url=base_url)
+    def test_overwrite_retry_count_header(
+        self, client: Openlayer, failures_before_success: int, respx_mock: MockRouter
+    ) -> None:
+        client = client.with_options(max_retries=4)
+
+        nb_retries = 0
+
+        def retry_handler(_request: httpx.Request) -> httpx.Response:
+            nonlocal nb_retries
+            if nb_retries < failures_before_success:
+                nb_retries += 1
+                return httpx.Response(500)
+            return httpx.Response(200)
+
+        respx_mock.post("/inference-pipelines/182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e/data-stream").mock(
+            side_effect=retry_handler
+        )
+
+        response = client.inference_pipelines.data.with_raw_response.stream(
+            inference_pipeline_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            config={"output_column_name": "output"},
+            rows=[
+                {
+                    "user_query": "bar",
+                    "output": "bar",
+                    "tokens": "bar",
+                    "cost": "bar",
+                    "timestamp": "bar",
+                }
+            ],
+            extra_headers={"x-stainless-retry-count": "42"},
+        )
+
+        assert response.http_request.headers.get("x-stainless-retry-count") == "42"
 
 
 class TestAsyncOpenlayer:
@@ -1486,6 +1574,7 @@ class Model(BaseModel):
             [3, "", 0.5],
             [2, "", 0.5 * 2.0],
             [1, "", 0.5 * 4.0],
+            [-1100, "", 7.8],  # test large number potentially overflowing
         ],
     )
     @mock.patch("time.time", mock.MagicMock(return_value=1696004797))
@@ -1524,7 +1613,7 @@ async def test_retrying_timeout_errors_doesnt_leak(self, respx_mock: MockRouter)
                                 "output": "42",
                                 "tokens": 7,
                                 "cost": 0.02,
-                                "timestamp": 1620000000,
+                                "timestamp": 1610000000,
                             }
                         ],
                     ),
@@ -1561,7 +1650,7 @@ async def test_retrying_status_errors_doesnt_leak(self, respx_mock: MockRouter)
                                 "output": "42",
                                 "tokens": 7,
                                 "cost": 0.02,
-                                "timestamp": 1620000000,
+                                "timestamp": 1610000000,
                             }
                         ],
                     ),
@@ -1576,8 +1665,13 @@ async def test_retrying_status_errors_doesnt_leak(self, respx_mock: MockRouter)
     @mock.patch("openlayer._base_client.BaseClient._calculate_retry_timeout", _low_retry_timeout)
     @pytest.mark.respx(base_url=base_url)
     @pytest.mark.asyncio
+    @pytest.mark.parametrize("failure_mode", ["status", "exception"])
     async def test_retries_taken(
-        self, async_client: AsyncOpenlayer, failures_before_success: int, respx_mock: MockRouter
+        self,
+        async_client: AsyncOpenlayer,
+        failures_before_success: int,
+        failure_mode: Literal["status", "exception"],
+        respx_mock: MockRouter,
     ) -> None:
         client = async_client.with_options(max_retries=4)
 
@@ -1587,6 +1681,8 @@ def retry_handler(_request: httpx.Request) -> httpx.Response:
             nonlocal nb_retries
             if nb_retries < failures_before_success:
                 nb_retries += 1
+                if failure_mode == "exception":
+                    raise RuntimeError("oops")
                 return httpx.Response(500)
             return httpx.Response(200)
 
@@ -1609,3 +1705,82 @@ def retry_handler(_request: httpx.Request) -> httpx.Response:
         )
 
         assert response.retries_taken == failures_before_success
+        assert int(response.http_request.headers.get("x-stainless-retry-count")) == failures_before_success
+
+    @pytest.mark.parametrize("failures_before_success", [0, 2, 4])
+    @mock.patch("openlayer._base_client.BaseClient._calculate_retry_timeout", _low_retry_timeout)
+    @pytest.mark.respx(base_url=base_url)
+    @pytest.mark.asyncio
+    async def test_omit_retry_count_header(
+        self, async_client: AsyncOpenlayer, failures_before_success: int, respx_mock: MockRouter
+    ) -> None:
+        client = async_client.with_options(max_retries=4)
+
+        nb_retries = 0
+
+        def retry_handler(_request: httpx.Request) -> httpx.Response:
+            nonlocal nb_retries
+            if nb_retries < failures_before_success:
+                nb_retries += 1
+                return httpx.Response(500)
+            return httpx.Response(200)
+
+        respx_mock.post("/inference-pipelines/182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e/data-stream").mock(
+            side_effect=retry_handler
+        )
+
+        response = await client.inference_pipelines.data.with_raw_response.stream(
+            inference_pipeline_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            config={"output_column_name": "output"},
+            rows=[
+                {
+                    "user_query": "bar",
+                    "output": "bar",
+                    "tokens": "bar",
+                    "cost": "bar",
+                    "timestamp": "bar",
+                }
+            ],
+            extra_headers={"x-stainless-retry-count": Omit()},
+        )
+
+        assert len(response.http_request.headers.get_list("x-stainless-retry-count")) == 0
+
+    @pytest.mark.parametrize("failures_before_success", [0, 2, 4])
+    @mock.patch("openlayer._base_client.BaseClient._calculate_retry_timeout", _low_retry_timeout)
+    @pytest.mark.respx(base_url=base_url)
+    @pytest.mark.asyncio
+    async def test_overwrite_retry_count_header(
+        self, async_client: AsyncOpenlayer, failures_before_success: int, respx_mock: MockRouter
+    ) -> None:
+        client = async_client.with_options(max_retries=4)
+
+        nb_retries = 0
+
+        def retry_handler(_request: httpx.Request) -> httpx.Response:
+            nonlocal nb_retries
+            if nb_retries < failures_before_success:
+                nb_retries += 1
+                return httpx.Response(500)
+            return httpx.Response(200)
+
+        respx_mock.post("/inference-pipelines/182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e/data-stream").mock(
+            side_effect=retry_handler
+        )
+
+        response = await client.inference_pipelines.data.with_raw_response.stream(
+            inference_pipeline_id="182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
+            config={"output_column_name": "output"},
+            rows=[
+                {
+                    "user_query": "bar",
+                    "output": "bar",
+                    "tokens": "bar",
+                    "cost": "bar",
+                    "timestamp": "bar",
+                }
+            ],
+            extra_headers={"x-stainless-retry-count": "42"},
+        )
+
+        assert response.http_request.headers.get("x-stainless-retry-count") == "42"
diff --git a/tests/test_models.py b/tests/test_models.py
index 963a34ff..f019e17b 100644
--- a/tests/test_models.py
+++ b/tests/test_models.py
@@ -245,7 +245,7 @@ class Model(BaseModel):
     assert m.foo is True
 
     m = Model.construct(foo="CARD_HOLDER")
-    assert m.foo is "CARD_HOLDER"
+    assert m.foo == "CARD_HOLDER"
 
     m = Model.construct(foo={"bar": False})
     assert isinstance(m.foo, Submodel1)
diff --git a/tests/test_response.py b/tests/test_response.py
index bc0a45bd..544ceeb4 100644
--- a/tests/test_response.py
+++ b/tests/test_response.py
@@ -190,6 +190,56 @@ async def test_async_response_parse_annotated_type(async_client: AsyncOpenlayer)
     assert obj.bar == 2
 
 
+@pytest.mark.parametrize(
+    "content, expected",
+    [
+        ("false", False),
+        ("true", True),
+        ("False", False),
+        ("True", True),
+        ("TrUe", True),
+        ("FalSe", False),
+    ],
+)
+def test_response_parse_bool(client: Openlayer, content: str, expected: bool) -> None:
+    response = APIResponse(
+        raw=httpx.Response(200, content=content),
+        client=client,
+        stream=False,
+        stream_cls=None,
+        cast_to=str,
+        options=FinalRequestOptions.construct(method="get", url="/foo"),
+    )
+
+    result = response.parse(to=bool)
+    assert result is expected
+
+
+@pytest.mark.parametrize(
+    "content, expected",
+    [
+        ("false", False),
+        ("true", True),
+        ("False", False),
+        ("True", True),
+        ("TrUe", True),
+        ("FalSe", False),
+    ],
+)
+async def test_async_response_parse_bool(client: AsyncOpenlayer, content: str, expected: bool) -> None:
+    response = AsyncAPIResponse(
+        raw=httpx.Response(200, content=content),
+        client=client,
+        stream=False,
+        stream_cls=None,
+        cast_to=str,
+        options=FinalRequestOptions.construct(method="get", url="/foo"),
+    )
+
+    result = await response.parse(to=bool)
+    assert result is expected
+
+
 class OtherModel(BaseModel):
     a: str