diff --git a/README.md b/README.md
index 8f4da7e9..9cd8b814 100644
--- a/README.md
+++ b/README.md
@@ -20,7 +20,12 @@ from log10.load import log10
 log10(openai)
 # all your openai calls are now logged
 ```
+For OpenAI v1, use `from log10.load import OpenAI` instead of `from openai import OpenAI`
+```python
+from log10.load import OpenAI
 
+client = OpenAI()
+```
 Access your LLM data at [log10.io](https://log10.io)
 
 
@@ -42,8 +47,12 @@ Log10 offers various integration methods, including a python LLM library wrapper
 Pick the one that works best for you.
 
 #### OpenAI
-Use library wrapper `log10(openai)` as shown [above](#-what-is-this).
-Full script [here](examples/logging/chatcompletion.py).
+| log10 ver| openai v0 | openai v1 |
+|----------|----------|----------|
+| 0.4 | `load(openai)` ✅ | ❌ |
+| 0.5 | `load(openai)` ✅ | `from log10.load import OpenAI` ✅ |
+
+**OpenAI v0** - Use library wrapper `log10(openai)`. Check out `examples/logging` in log10 version `0.4.6`.
 ```python
 import openai
 from log10.load import log10
@@ -52,12 +61,27 @@ log10(openai)
 # openai calls are now logged
 ```
 
-Use Log10 LLM abstraction.
-Full script [here](examples/logging/llm_abstraction.py#6-#14).
+**OpenAI v1**
+> NOTE: We added OpenAI v1 API support in log10 `0.5.0` release. `load.log10(openai)` still works for openai v1.
+```python
+from log10.load import OpenAI
+# from openai import OpenAI
+
+client = OpenAI()
+completion = client.completions.create(model="curie", prompt="Once upon a time")
+# All completions.create and chat.completions.create calls will be logged
+```
+Full script [here](examples/logging/completion.py).
+
+
+**Use Log10 LLM abstraction**
+
 ```python
 from log10.openai import OpenAI
+
 llm = OpenAI({"model": "gpt-3.5-turbo"}, log10_config=Log10Config())
 ```
+openai v1+ lib required. Full script [here](examples/logging/llm_abstraction.py#6-#14).
 
 #### Anthropic
 Use library wrapper `log10(anthropic)`.
@@ -74,6 +98,7 @@ Use Log10 LLM abstraction.
 Full script [here](examples/logging/llm_abstraction.py#16-#19).
 ```python
 from log10.anthropic import Anthropic
+
 llm = Anthropic({"model": "claude-2"}, log10_config=Log10Config())
 ```
 
@@ -85,12 +110,14 @@ Adding other providers is on the roadmap.
 **MosaicML** with LLM abstraction. Full script [here](/examples/logging/mosaicml_completion.py).
 ```python
 from log10.mosaicml import MosaicML
+
 llm = MosaicML({"model": "llama2-70b-chat/v1"}, log10_config=Log10Config())
 ```
 
 **Together** with LLM abstraction. Full script [here](/examples/logging/together_completion.py).
 ```python
 from log10.together import Together
+
 llm = Together({"model": "togethercomputer/llama-2-70b-chat"}, log10_config=Log10Config())
 ```
 
@@ -137,7 +164,7 @@ And provide the following configuration in either a `.env` file, or as environme
 
 ### 🧠🔁 Readiness for RLHF & self hosting
 
-Use your data and feedback from users to fine-tune custom models with RLHF with the option of building and deploying more reliable, accurate and efficient self-hosted models. 
+Use your data and feedback from users to fine-tune custom models with RLHF with the option of building and deploying more reliable, accurate and efficient self-hosted models.
 
 ### 👥🤝 Collaboration
 
@@ -158,7 +185,7 @@ Create flexible groups to share and collaborate over all of the above features
 You can find and run examples under folder `examples`, e.g. run a logging example:
 ```
 python examples/logging/chatcompletion.py
-``` 
+```
 
 Also you can run some end-to-end tests with [`xdocttest`](https://github.com/Erotemic/xdoctest) installed (`pip install xdoctest`).
 
diff --git a/examples/README.md b/examples/README.md
index 08b16064..8e36c35c 100644
--- a/examples/README.md
+++ b/examples/README.md
@@ -25,6 +25,7 @@ Can be run on `OpenAI` or `Anthropic`
 ## Logging (and debugging)
 
 ### OpenAI
+Requires openai >= "1.0.0"
 
 - `chatCompletion_async_vs_sync.py` Compare latencies when logging in async vs sync mode
 - `chatCompletion.py` Chat endpoint example
diff --git a/examples/logging/chatcompletion.py b/examples/logging/chatcompletion.py
index 70fbd8b3..521503c9 100644
--- a/examples/logging/chatcompletion.py
+++ b/examples/logging/chatcompletion.py
@@ -1,15 +1,9 @@
-import os
+from log10.load import OpenAI
 
-import openai
 
-from log10.load import log10
+client = OpenAI()
 
-
-log10(openai)
-
-openai.api_key = os.getenv("OPENAI_API_KEY")
-
-completion = openai.ChatCompletion.create(
+completion = client.chat.completions.create(
     model="gpt-3.5-turbo",
     messages=[
         {
diff --git a/examples/logging/chatcompletion_async_vs_sync.py b/examples/logging/chatcompletion_async_vs_sync.py
deleted file mode 100644
index c52ffda8..00000000
--- a/examples/logging/chatcompletion_async_vs_sync.py
+++ /dev/null
@@ -1,70 +0,0 @@
-import sys
-
-
-if "init_modules" in globals():
-    # second or subsequent run: remove all but initially loaded modules
-    for m in list(sys.modules.keys()):
-        if m not in init_modules:
-            del sys.modules[m]
-else:
-    # first run: find out which modules were initially loaded
-    init_modules = list(sys.modules.keys())
-
-import os
-
-import openai
-
-from log10.load import log10, log10_session
-
-
-openai.api_key = os.getenv("OPENAI_API_KEY")
-
-# Launch an async run
-with log10_session():
-    log10(openai, DEBUG_=True, USE_ASYNC_=True)
-    completion = openai.ChatCompletion.create(
-        model="gpt-3.5-turbo",
-        messages=[
-            {
-                "role": "system",
-                "content": "You are the most knowledgable Star Wars guru on the planet",
-            },
-            {
-                "role": "user",
-                "content": "Write the time period of all the Star Wars movies and spinoffs?",
-            },
-        ],
-    )
-    print(completion.choices[0].message)
-
-# reload modules to prevent double calling openAI
-if "init_modules" in globals():
-    # second or subsequent run: remove all but initially loaded modules
-    for m in list(sys.modules.keys()):
-        if m not in init_modules:
-            del sys.modules[m]
-else:
-    # first run: find out which modules were initially loaded
-    init_modules = list(sys.modules.keys())
-
-
-import openai  # noqa
-
-# Compare to sync run - note there can be variability in the OpenAI calls
-with log10_session():
-    log10(openai, DEBUG_=True, USE_ASYNC_=False)
-
-    completion = openai.ChatCompletion.create(
-        model="gpt-3.5-turbo",
-        messages=[
-            {
-                "role": "system",
-                "content": "You are the most knowledgable Star Wars guru on the planet",
-            },
-            {
-                "role": "user",
-                "content": "Write the time period of all the Star Wars movies and spinoffs?",
-            },
-        ],
-    )
-    print(completion.choices[0].message)
diff --git a/examples/logging/chatcompletion_sync.py b/examples/logging/chatcompletion_sync.py
new file mode 100644
index 00000000..ffe97e30
--- /dev/null
+++ b/examples/logging/chatcompletion_sync.py
@@ -0,0 +1,27 @@
+from log10.load import OpenAI, log10_session
+
+
+client = OpenAI()
+
+with log10_session(tags=["log10-io/examples"]):
+    completion = client.chat.completions.create(
+        model="gpt-3.5-turbo",
+        messages=[
+            {
+                "role": "user",
+                "content": "Hello?",
+            },
+        ],
+    )
+    print(completion.choices[0].message)
+
+    completion = client.chat.completions.create(
+        model="gpt-3.5-turbo",
+        messages=[
+            {
+                "role": "user",
+                "content": "Hello again, are you there?",
+            },
+        ],
+    )
+    print(completion.choices[0].message)
diff --git a/examples/logging/completion.py b/examples/logging/completion.py
index ff1c23fc..5007d118 100644
--- a/examples/logging/completion.py
+++ b/examples/logging/completion.py
@@ -1,15 +1,9 @@
-import os
+from log10.load import OpenAI
 
-import openai
 
-from log10.load import log10
+client = OpenAI()
 
-
-log10(openai)
-
-openai.api_key = os.getenv("OPENAI_API_KEY")
-
-response = openai.Completion.create(
+response = client.completions.create(
     model="gpt-3.5-turbo-instruct",
     prompt="Write the names of all Star Wars movies and spinoffs along with the time periods in which they were set?",
     temperature=0,
diff --git a/examples/logging/completion_simple.py b/examples/logging/completion_simple.py
index beae1fbf..b293df2f 100644
--- a/examples/logging/completion_simple.py
+++ b/examples/logging/completion_simple.py
@@ -1,15 +1,9 @@
-import os
+from log10.load import OpenAI
 
-import openai
 
-from log10.load import log10
+client = OpenAI()
 
-
-log10(openai, DEBUG_=True, USE_ASYNC_=False)
-
-openai.api_key = os.getenv("OPENAI_API_KEY")
-
-response = openai.Completion.create(
+response = client.completions.create(
     model="gpt-3.5-turbo-instruct",
     prompt="What is 2+2?",
     temperature=0,
diff --git a/examples/logging/get_url.py b/examples/logging/get_url.py
index d72b7df5..1ef2cae8 100644
--- a/examples/logging/get_url.py
+++ b/examples/logging/get_url.py
@@ -1,21 +1,12 @@
-import os
+from log10.load import OpenAI, log10_session
 
-import openai
-from langchain.llms import OpenAI
 
-from log10.load import log10, log10_session
-
-
-log10(openai)
-
-openai.api_key = os.getenv("OPENAI_API_KEY")
-
-llm = OpenAI(temperature=0.9, model_name="gpt-3.5-turbo-instruct")
+client = OpenAI()
 
 with log10_session() as session:
     print(session.last_completion_url())
 
-    response = openai.Completion.create(
+    response = client.completions.create(
         model="gpt-3.5-turbo-instruct",
         prompt="Why did the chicken cross the road?",
         temperature=0,
@@ -27,7 +18,7 @@
 
     print(session.last_completion_url())
 
-    response = openai.Completion.create(
+    response = client.completions.create(
         model="gpt-3.5-turbo-instruct",
         prompt="Why did the cow cross the road?",
         temperature=0,
@@ -42,7 +33,7 @@
 with log10_session() as session:
     print(session.last_completion_url())
 
-    response = openai.Completion.create(
+    response = client.completions.create(
         model="gpt-3.5-turbo-instruct",
         prompt="Why did the frog cross the road?",
         temperature=0,
@@ -54,7 +45,7 @@
 
     print(session.last_completion_url())
 
-    response = openai.Completion.create(
+    response = client.completions.create(
         model="gpt-3.5-turbo-instruct",
         prompt="Why did the scorpion cross the road?",
         temperature=0,
diff --git a/examples/logging/langchain_multiple_tools.py b/examples/logging/langchain_multiple_tools.py
index 7be2a8cd..d38b9003 100644
--- a/examples/logging/langchain_multiple_tools.py
+++ b/examples/logging/langchain_multiple_tools.py
@@ -1,5 +1,3 @@
-import os
-
 import openai
 
 from log10.load import log10
@@ -7,7 +5,6 @@
 
 log10(openai)
 
-openai.api_key = os.getenv("OPENAI_API_KEY")
 MAX_TOKENS = 512
 TOOLS_DEFAULT_LIST = ["llm-math", "wikipedia"]
 
diff --git a/examples/logging/langchain_qa.py b/examples/logging/langchain_qa.py
index afe814d9..8cc027e9 100644
--- a/examples/logging/langchain_qa.py
+++ b/examples/logging/langchain_qa.py
@@ -1,5 +1,3 @@
-import os
-
 import openai
 
 from log10.load import log10
@@ -7,8 +5,6 @@
 
 log10(openai)
 
-openai.api_key = os.getenv("OPENAI_API_KEY")
-
 # Example from: https://python.langchain.com/en/latest/use_cases/question_answering.html
 # Download the state_of_the_union.txt here: https://raw.githubusercontent.com/hwchase17/langchain/master/docs/modules/state_of_the_union.txt
 # This example requires: pip install chromadb
diff --git a/examples/logging/langchain_simple_sequential.py b/examples/logging/langchain_simple_sequential.py
index b94f5abb..24bcf219 100644
--- a/examples/logging/langchain_simple_sequential.py
+++ b/examples/logging/langchain_simple_sequential.py
@@ -1,5 +1,3 @@
-import os
-
 import openai
 
 from log10.load import log10
@@ -7,8 +5,6 @@
 
 log10(openai)
 
-openai.api_key = os.getenv("OPENAI_API_KEY")
-
 from langchain.chains import LLMChain, SimpleSequentialChain
 from langchain.llms import OpenAI
 from langchain.prompts import PromptTemplate
diff --git a/examples/logging/langchain_sqlagent.py b/examples/logging/langchain_sqlagent.py
index 28a3583a..fa4c6a38 100644
--- a/examples/logging/langchain_sqlagent.py
+++ b/examples/logging/langchain_sqlagent.py
@@ -1,5 +1,4 @@
 import datetime
-import os
 import random
 
 import openai
@@ -78,9 +77,6 @@ def generate_random_user():
 
 session.close()
 
-# Setup vars for Langchain
-openai.api_key = os.getenv("OPENAI_API_KEY")
-
 # Setup Langchain SQL agent
 db = SQLDatabase.from_uri("sqlite:///users.db")
 toolkit = SQLDatabaseToolkit(db=db)
diff --git a/examples/logging/long_context_exception.py b/examples/logging/long_context_exception.py
index d0253321..cff79206 100644
--- a/examples/logging/long_context_exception.py
+++ b/examples/logging/long_context_exception.py
@@ -1,17 +1,11 @@
-import os
+from log10.load import OpenAI
 
-import openai
 
-from log10.load import log10
-
-
-log10(openai)
-
-openai.api_key = os.getenv("OPENAI_API_KEY")
+client = OpenAI()
 
 text_to_repeat = "What is the meaning of life?" * 1000
 
-response = openai.Completion.create(
+response = client.completions.create(
     model="gpt-3.5-turbo-instruct",
     prompt=text_to_repeat,
     temperature=0,
diff --git a/examples/logging/multiple_sessions.py b/examples/logging/multiple_sessions.py
index d60f2192..fc61e2ef 100644
--- a/examples/logging/multiple_sessions.py
+++ b/examples/logging/multiple_sessions.py
@@ -1,5 +1,3 @@
-import os
-
 import openai
 from langchain.chains import LLMChain, SimpleSequentialChain
 from langchain.llms import OpenAI
@@ -10,8 +8,6 @@
 
 log10(openai)
 
-openai.api_key = os.getenv("OPENAI_API_KEY")
-
 
 llm = OpenAI(temperature=0.9, model_name="gpt-3.5-turbo-instruct")
 
diff --git a/examples/logging/tags_mixed.py b/examples/logging/tags_mixed.py
index c46f5ae9..dfcae10a 100644
--- a/examples/logging/tags_mixed.py
+++ b/examples/logging/tags_mixed.py
@@ -6,8 +6,9 @@
 
 log10(openai)
 
+client = openai.OpenAI()
 with log10_session(tags=["foo", "bar"]):
-    response = openai.Completion.create(
+    response = client.completions.create(
         model="gpt-3.5-turbo-instruct",
         prompt="Where is the Eiffel Tower?",
         temperature=0,
diff --git a/examples/logging/tags_openai.py b/examples/logging/tags_openai.py
index c14de75e..4e5bac88 100644
--- a/examples/logging/tags_openai.py
+++ b/examples/logging/tags_openai.py
@@ -1,16 +1,8 @@
-import os
+from log10.load import OpenAI, log10_session
 
-import openai
 
-from log10.load import log10, log10_session
-
-
-log10(openai)
-
-openai.api_key = os.getenv("OPENAI_API_KEY")
-
-
-response = openai.Completion.create(
+client = OpenAI()
+response = client.completions.create(
     model="gpt-3.5-turbo-instruct",
     prompt="Where are the pyramids?",
     temperature=0,
@@ -22,7 +14,7 @@
 print(response)
 
 with log10_session(tags=["foo", "bar"]):
-    response = openai.Completion.create(
+    response = client.completions.create(
         model="gpt-3.5-turbo-instruct",
         prompt="Where is the Eiffel Tower?",
         temperature=0,
@@ -34,7 +26,7 @@
     print(response)
 
 with log10_session(tags=["bar", "baz"]):
-    response = openai.Completion.create(
+    response = client.completions.create(
         model="gpt-3.5-turbo-instruct",
         prompt="Where is the statue of liberty?",
         temperature=0,
@@ -45,7 +37,7 @@
     )
     print(response)
 
-response = openai.Completion.create(
+response = client.completions.create(
     model="gpt-3.5-turbo-instruct",
     prompt="Where is machu picchu?",
     temperature=0,
diff --git a/log10/load.py b/log10/load.py
index 31b5f072..475655be 100644
--- a/log10/load.py
+++ b/log10/load.py
@@ -1,6 +1,5 @@
 import asyncio
 import functools
-import inspect
 import json
 import logging
 import os
@@ -9,12 +8,12 @@
 import time
 import traceback
 from contextlib import contextmanager
+from importlib.metadata import version
 
-import backoff  # for exponential backoff
+import backoff
 import requests
 from dotenv import load_dotenv
-
-from log10.openai import RETRY_ERROR_TYPES as OPENAI_RETRY_ERROR_TYPES
+from packaging.version import parse
 
 
 load_dotenv()
@@ -42,9 +41,31 @@
     target_service = "log10"  # default to log10
 
 
-@backoff.on_exception(backoff.expo, OPENAI_RETRY_ERROR_TYPES)
+def is_openai_v1() -> bool:
+    """Return whether OpenAI API is v1 or more."""
+    _version = parse(version("openai"))
+    return _version.major >= 1
+
+
 def func_with_backoff(func, *args, **kwargs):
-    return func(*args, **kwargs)
+    if func.__module__ != "openai" or is_openai_v1():
+        return func(*args, **kwargs)
+
+    import openai
+
+    retry_errors = (
+        openai.error.APIConnectionError,
+        openai.error.APIError,
+        openai.error.RateLimitError,
+        openai.error.ServiceUnavailableError,
+        openai.error.Timeout,
+    )
+
+    @backoff.on_exception(backoff.expo, retry_errors)
+    def _func_with_backoff(func, *args, **kwargs):
+        return func(*args, **kwargs)
+
+    return _func_with_backoff(func, *args, **kwargs)
 
 
 # todo: should we do backoff as well?
@@ -306,19 +327,25 @@ def wrapper(*args, **kwargs):
                     from log10.anthropic import Anthropic
 
                     response = Anthropic.prepare_response(kwargs["prompt"], output, "text")
+                    kind = "completion"
                 else:
                     response = output
+                    kind = "chat" if output.object == "chat.completion" else "completion"
 
                 # in case the usage of load(openai) and langchain.ChatOpenAI
                 if "api_key" in kwargs:
                     kwargs.pop("api_key")
 
+                if hasattr(response, "model_dump_json"):
+                    response = response.model_dump_json()
+                else:
+                    response = json.dumps(response)
                 log_row = {
-                    "response": json.dumps(response),
+                    "response": response,
                     "status": "finished",
                     "duration": int(duration * 1000),
                     "stacktrace": json.dumps(stacktrace),
-                    "kind": "completion",
+                    "kind": kind,
                     "orig_module": func.__module__,
                     "orig_qualname": func.__qualname__,
                     "request": json.dumps(kwargs),
@@ -361,42 +388,35 @@ def set_sync_log_text(USE_ASYNC=True):
 
 def log10(module, DEBUG_=False, USE_ASYNC_=True):
     """Intercept and overload module for logging purposes
+    support both openai V0 and V1, and anthropic
 
     Keyword arguments:
     module -- the module to be intercepted (e.g. openai)
     DEBUG_ -- whether to show log10 related debug statements via python logging (default False)
     USE_ASYNC_ -- whether to run in async mode (default True)
 
+    Openai V0 example:
     Example:
-        >>> from log10.load import log10
+        >>> from log10.load import log10 # xdoctest: +SKIP
         >>> import openai
         >>> log10(openai)
-        >>> response = openai.Completion.create(
+        >>> completion = openai.Completion.create(
         >>>     model="gpt-3.5-turbo-instruct",
         >>>     prompt="Once upon a time",
         >>>     max_tokens=32,
         >>> )
-        >>> print(response)
+        >>> print(completion)
 
     Example:
-        >>> from log10.load import log10
+        >>> from log10.load import log10 # xdoctest: +SKIP
         >>> import openai
         >>> log10(openai)
-        >>> response = openai.ChatCompletion.create(
+        >>> completion = openai.ChatCompletion.create(
         >>>     model="gpt-3.5-turbo",
-        >>>     messages=[
-        >>>         {
-        >>>             "role": "system",
-        >>>             "content": "You are a Pingpong machine.",
-        >>>         },
-        >>>         {
-        >>>             "role": "user",
-        >>>             "content": "Ping",
-        >>>         },
-        >>>     ],
+        >>>     messages=[{"role": "user", "content": "Hello world"}],
         >>>     max_tokens=8,
         >>> )
-        >>> print(response)
+        >>> print(completion)
 
     Example:
         >>> from log10.load import log10
@@ -467,24 +487,32 @@ def log10(module, DEBUG_=False, USE_ASYNC_=True):
     #         elif inspect.isclass(method):  # Handle nested classes
     #             intercept_class_methods(method)
 
-    for name, attr in vars(module).items():
-        if inspect.isclass(attr):
-            # OpenAI
-            if module.__name__ == "openai" and name in ["ChatCompletion", "Completion"]:
-                for method_name, method in vars(attr).items():
-                    if isinstance(method, classmethod):
-                        original_method = method.__func__
-                        if original_method.__qualname__ in [
-                            "ChatCompletion.create",
-                            "Completion.create",
-                        ]:
-                            decorated_method = intercepting_decorator(original_method)
-                            setattr(attr, method_name, classmethod(decorated_method))
-    # Anthropic
     if module.__name__ == "anthropic":
-        method = getattr(module.resources.completions.Completions, "create")
         attr = module.resources.completions.Completions
+        method = getattr(attr, "create")
         setattr(attr, "create", intercepting_decorator(method))
+    elif module.__name__ == "openai":
+        openai_version = parse(version("openai"))
+        global OPENAI_V1
+        OPENAI_V1 = openai_version >= parse("1.0.0")
+
+        # support for sync completions
+        if OPENAI_V1:
+            attr = module.resources.completions.Completions
+            method = getattr(attr, "create")
+            setattr(attr, "create", intercepting_decorator(method))
+
+            attr = module.resources.chat.completions.Completions
+            method = getattr(attr, "create")
+            setattr(attr, "create", intercepting_decorator(method))
+        else:
+            attr = module.api_resources.completion.Completion
+            method = getattr(attr, "create")
+            setattr(attr, "create", intercepting_decorator(method))
+
+            attr = module.api_resources.chat_completion.ChatCompletion
+            mothod = getattr(attr, "create")
+            setattr(attr, "create", intercepting_decorator(mothod))
 
         # For future reference:
         # if callable(attr) and isinstance(attr, types.FunctionType):
@@ -494,3 +522,37 @@ def log10(module, DEBUG_=False, USE_ASYNC_=True):
         #     intercept_class_methods(attr)
         # # else: # uncomment if we want to include nested function support
         # #     intercept_nested_functions(attr)
+
+
+if is_openai_v1():
+    import openai
+    from openai import OpenAI
+
+    class OpenAI(OpenAI):
+        """
+        Example:
+            >>> from log10.load import OpenAI
+            >>> client = OpenAI(tags=["load_v1_test"])
+            >>> completion = client.completions.create(model='curie', prompt="Twice upon a time", max_tokens=32)
+            >>> print(completion)
+
+        Example:
+            >>> from log10.load import OpenAI
+            >>> client = OpenAI(tags=["load_v1_test"])
+            >>> completion = client.chat.completions.create(
+            >>>     model="gpt-3.5-turbo",
+            >>>     messages=[{"role": "user", "content": "Hello world"}],
+            >>> )
+            >>> print(completion)
+        """
+
+        def __init__(self, *args, **kwargs):
+            # check if tags is passed in
+            if "tags" in kwargs:
+                global global_tags
+                global_tags = kwargs.pop("tags")
+            super().__init__(*args, **kwargs)
+
+            if not getattr(openai, "_log10_patched", False):
+                log10(openai)
+                openai._log10_patched = True
diff --git a/log10/openai.py b/log10/openai.py
index 88b1f7fb..d30c8a06 100644
--- a/log10/openai.py
+++ b/log10/openai.py
@@ -1,46 +1,51 @@
 import time
 from copy import deepcopy
+from importlib.metadata import version
 from typing import List
 
-# for exponential backoff
-import backoff
 import openai
-from openai import error
+from packaging.version import parse
 
 from log10.llm import LLM, ChatCompletion, Kind, Message, TextCompletion
 
 
-RETRY_ERROR_TYPES = (
-    error.APIConnectionError,
-    error.APIError,
-    error.RateLimitError,
-    error.ServiceUnavailableError,
-    error.Timeout,
-)
+def is_openai_v1() -> bool:
+    """Return whether OpenAI API is v1 or more."""
+    _version = parse(version("openai"))
+    return _version.major >= 1
 
 
 class OpenAI(LLM):
     def __init__(self, hparams: dict = None, log10_config=None):
+        assert is_openai_v1(), "OpenAI API version must be >= 1.0.0 to use log10.OpenAI class"
         super().__init__(hparams, log10_config)
+        self._client = openai.OpenAI()
 
-    @backoff.on_exception(backoff.expo, RETRY_ERROR_TYPES)
     def chat(self, messages: List[Message], hparams: dict = None) -> ChatCompletion:
+        """
+        Example:
+            >>> from log10.llm import Log10Config, Message
+            >>> from log10.openai import OpenAI
+            >>> llm = OpenAI({"model": "gpt-3.5-turbo"}, log10_config=Log10Config())
+            >>> response = llm.chat([Message(role="user", content="Hello")])
+            >>> print(response)
+        """
         request = self.chat_request(messages, hparams)
 
         start_time = time.perf_counter()
-        completion = openai.ChatCompletion.create(**request)
+        completion = self._client.chat.completions.create(**request)
 
         self.completion_id = self.log_start(request, Kind.chat)
 
         response = ChatCompletion(
-            role=completion.choices[0]["message"]["role"],
-            content=completion.choices[0]["message"]["content"],
+            role=completion.choices[0].message.role,
+            content=completion.choices[0].message.content,
             response=completion,
         )
 
         self.log_end(
             self.completion_id,
-            completion,
+            completion.model_dump(),
             time.perf_counter() - start_time,
         )
 
@@ -57,19 +62,26 @@ def chat_request(self, messages: List[Message], hparams: dict = None) -> dict:
             **merged_hparams,
         }
 
-    @backoff.on_exception(backoff.expo, RETRY_ERROR_TYPES)
     def text(self, prompt: str, hparams: dict = None) -> TextCompletion:
+        """
+        Example:
+            >>> from log10.llm import Log10Config
+            >>> from log10.openai import OpenAI
+            >>> llm = OpenAI({"model": "gpt-3.5-turbo-instruct"}, log10_config=Log10Config())
+            >>> response = llm.text("This is a test.")
+            >>> print(response)
+        """
         request = self.text_request(prompt, hparams)
 
         start_time = time.perf_counter()
         completion_id = self.log_start(request, Kind.text)
 
-        completion = openai.Completion.create(**request)
+        completion = self._client.completions.create(**request)
         response = TextCompletion(text=completion.choices[0].text, response=completion)
 
         self.log_end(
             completion_id,
-            completion,
+            completion.model_dump(),
             time.perf_counter() - start_time,
         )
 
diff --git a/logging.md b/logging.md
index e12e1ed0..9a6b6dea 100644
--- a/logging.md
+++ b/logging.md
@@ -23,6 +23,18 @@ from log10.load import log10
 
 log10(openai)
 ```
+This works for both openai v0 and v1.
+
+For openai v1 only, the logging can be done like this:
+
+```python
+from log10.load import OpenAI
+# from openai import OpenAI
+
+client = OpenAI()
+completion = client.completions.create(model='curie', prompt="Once upon a time")
+# All completions.create and chat.completions.create calls will be logged
+```
 
 From that point on, all openai calls during the process execution will be logged.
 Please note that this only works for non-streaming calls at this time.
@@ -125,6 +137,12 @@ with log10_session(tags=["foo", "bar"]):
 
 ```
 
+An addition way to add tag with openai v1 API:
+```python
+from log10.load import OpenAI
+client = OpenAI(tags=["foo"])
+```
+
 ## LangChain logger
 
 If you want to use LLMs which are not natively supported by log10 yet, you can use the log10 logger / callback with langchain's llm abstraction:
diff --git a/pyproject.toml b/pyproject.toml
index ab9efb6a..1f65ee75 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -5,7 +5,7 @@ build-backend = "poetry.core.masonry.api"
 [tool.poetry]
 name = "log10-io"
 
-version = "0.4.6"
+version = "0.5.0"
 authors = ["log10 team"]
 license = "MIT"
 description = "Unified LLM data management"
@@ -29,7 +29,7 @@ build = "^0.10.0"
 [tool.poetry.dependencies]
 python = ">=3.9,<4.0"
 langchain = "^0.0.303"
-openai = "^0.28.1"
+openai = "<2"
 requests = "^2.31.0"
 python-dotenv = "^1.0.0"
 google-cloud-bigquery = "^3.11.4"
diff --git a/setup.py b/setup.py
index 045754d1..60f810a3 100644
--- a/setup.py
+++ b/setup.py
@@ -5,7 +5,7 @@
 
 setup(
     name="Log10",
-    version="0.4.6",
+    version="0.5.0",
     description="Log10 LLM data management",
     author="Log10 team",
     author_email="team@log10.io",