diff --git a/cogs/commands.py b/cogs/commands.py
index ef48595..2e43229 100644
--- a/cogs/commands.py
+++ b/cogs/commands.py
@@ -301,6 +301,11 @@ class Commands(discord.Cog, name="Commands"):
     @discord.option(
         name="prompt", description="The prompt to send to GPT3", required=True
     )
+    @discord.option(
+        name="private", 
+        description="Will only be visible to you", 
+        required=False
+    )
     @discord.option(
         name="temperature",
         description="Higher values means the model will take more risks",
@@ -334,13 +339,14 @@ class Commands(discord.Cog, name="Commands"):
         self,
         ctx: discord.ApplicationContext,
         prompt: str,
+        private: bool,
         temperature: float,
         top_p: float,
         frequency_penalty: float,
         presence_penalty: float,
     ):
         await self.converser_cog.ask_command(
-            ctx, prompt, temperature, top_p, frequency_penalty, presence_penalty
+            ctx, prompt, private, temperature, top_p, frequency_penalty, presence_penalty
         )
 
     @add_to_group("gpt")
@@ -360,6 +366,11 @@ class Commands(discord.Cog, name="Commands"):
         required=False,
         default="",
     )
+    @discord.option(
+        name="private", 
+        description="Will only be visible to you", 
+        required=False
+    )
     @discord.option(
         name="temperature",
         description="Higher values means the model will take more risks",
@@ -385,12 +396,13 @@ class Commands(discord.Cog, name="Commands"):
         ctx: discord.ApplicationContext,
         instruction: str,
         text: str,
+        private: bool,
         temperature: float,
         top_p: float,
         codex: bool,
     ):
         await self.converser_cog.edit_command(
-            ctx, instruction, text, temperature, top_p, codex
+            ctx, instruction, text, private, temperature, top_p, codex
         )
 
     @add_to_group("gpt")
@@ -422,6 +434,13 @@ class Commands(discord.Cog, name="Commands"):
         required=False,
         default=False,
     )
+    @discord.option(
+        name="model",
+        description="Which model to use with the bot",
+        required=False,
+        default=False,
+        autocomplete=Settings_autocompleter.get_models,
+    )
     @discord.option(
         name="temperature",
         description="Higher values means the model will take more risks",
@@ -462,6 +481,7 @@ class Commands(discord.Cog, name="Commands"):
         opener_file: str,
         private: bool,
         minimal: bool,
+        model: str,
         temperature: float,
         top_p: float,
         frequency_penalty: float,
@@ -473,6 +493,7 @@ class Commands(discord.Cog, name="Commands"):
             opener_file,
             private,
             minimal,
+            model,
             temperature,
             top_p,
             frequency_penalty,
diff --git a/cogs/text_service_cog.py b/cogs/text_service_cog.py
index 88fb3bb..8f01b84 100644
--- a/cogs/text_service_cog.py
+++ b/cogs/text_service_cog.py
@@ -693,6 +693,7 @@ class GPT3ComCon(discord.Cog, name="GPT3ComCon"):
         self,
         ctx: discord.ApplicationContext,
         prompt: str,
+        private: bool,
         temperature: float,
         top_p: float,
         frequency_penalty: float,
@@ -720,7 +721,7 @@ class GPT3ComCon(discord.Cog, name="GPT3ComCon"):
             if not user_api_key:
                 return
 
-        await ctx.defer()
+        await ctx.defer(ephemeral=private)
 
         overrides = Override(temperature, top_p, frequency_penalty, presence_penalty)
 
@@ -741,6 +742,7 @@ class GPT3ComCon(discord.Cog, name="GPT3ComCon"):
         ctx: discord.ApplicationContext,
         instruction: str,
         text: str,
+        private: bool,
         temperature: float,
         top_p: float,
         codex: bool,
@@ -766,7 +768,7 @@ class GPT3ComCon(discord.Cog, name="GPT3ComCon"):
             if not user_api_key:
                 return
 
-        await ctx.defer()
+        await ctx.defer(ephemeral=private)
 
         overrides = Override(temperature, top_p, 0, 0)
 
@@ -801,6 +803,7 @@ class GPT3ComCon(discord.Cog, name="GPT3ComCon"):
         opener_file: str,
         private: bool,
         minimal: bool,
+        model: str,
         temperature: float,
         top_p: float,
         frequency_penalty: float,
@@ -814,6 +817,7 @@ class GPT3ComCon(discord.Cog, name="GPT3ComCon"):
             opener_file (str): A .txt or .json file which is appended before the opener
             private (bool): If the thread should be private
             minimal (bool): If a minimal starter should be used
+            model (str): The openai model that should be used
             temperature (float): Sets the temperature override
             top_p (float): Sets the top p override
             frequency_penalty (float): Sets the frequency penalty override
@@ -866,7 +870,7 @@ class GPT3ComCon(discord.Cog, name="GPT3ComCon"):
             )
 
         self.conversation_threads[thread.id] = Thread(thread.id)
-        self.conversation_threads[thread.id].model = self.model.model
+        self.conversation_threads[thread.id].model = self.model.model if not model else model
 
         # Set the overrides for the conversation
         self.conversation_threads[thread.id].set_overrides(
diff --git a/gpt3discord.py b/gpt3discord.py
index 5068831..3ccd616 100644
--- a/gpt3discord.py
+++ b/gpt3discord.py
@@ -30,7 +30,8 @@ from services.environment_service import EnvService
 from models.openai_model import Model
 
 
-__version__ = "9.0.4"
+__version__ = "9.1"
+
 
 
 PID_FILE = Path("bot.pid")
diff --git a/models/autocomplete_model.py b/models/autocomplete_model.py
index 8525cb8..c89acfd 100644
--- a/models/autocomplete_model.py
+++ b/models/autocomplete_model.py
@@ -86,6 +86,17 @@ class Settings_autocompleter:
         await ctx.interaction.response.defer()  # defer so the autocomplete in int values doesn't error but rather just says not found
         return []
 
+    async def get_models(
+        ctx: discord.AutocompleteContext,
+    ):
+        """Gets all models"""
+        return [
+            value
+            for value in Models.TEXT_MODELS
+            if value.startswith(ctx.value.lower())
+        ]
+    
+    
     async def get_value_moderations(
         ctx: discord.AutocompleteContext,
     ):  # Behaves a bit weird if you go back and edit the parameter without typing in a new command
diff --git a/models/openai_model.py b/models/openai_model.py
index fd0d18f..ff65a5c 100644
--- a/models/openai_model.py
+++ b/models/openai_model.py
@@ -78,7 +78,7 @@ class Models:
 
     @staticmethod
     def get_max_tokens(model: str) -> int:
-        return Models.TOKEN_MAPPING.get(model, 4024)
+        return Models.TOKEN_MAPPING.get(model, 2024)
 
 
 class ImageSize:
@@ -576,11 +576,16 @@ class Model:
         self._prompt_min_length = value
         SETTINGS_DB["prompt_min_length"] = value
 
-    def backoff_handler(details):
+    def backoff_handler_http(details):
         print(
             f"Backing off {details['wait']:0.1f} seconds after {details['tries']} tries calling function {details['target']} | "
             f"{details['exception'].status}: {details['exception'].message}"
         )
+    def backoff_handler_request(details):
+        print(
+            f"Backing off {details['wait']:0.1f} seconds after {details['tries']} tries calling function {details['target']} | "
+            f"{details['exception'].args[0]}"
+        )
 
     async def valid_text_request(self, response):
         try:
@@ -598,7 +603,7 @@ class Model:
         factor=3,
         base=5,
         max_tries=4,
-        on_backoff=backoff_handler,
+        on_backoff=backoff_handler_http,
     )
     async def send_embedding_request(self, text, custom_api_key=None):
         async with aiohttp.ClientSession(raise_for_status=True) as session:
@@ -624,11 +629,11 @@ class Model:
 
     @backoff.on_exception(
         backoff.expo,
-        aiohttp.ClientResponseError,
+        ValueError,
         factor=3,
         base=5,
-        max_tries=6,
-        on_backoff=backoff_handler,
+        max_tries=4,
+        on_backoff=backoff_handler_request,
     )
     async def send_edit_request(
         self,
@@ -651,7 +656,7 @@ class Model:
         )
         print(f"Overrides -> temp:{temp_override}, top_p:{top_p_override}")
 
-        async with aiohttp.ClientSession(raise_for_status=True) as session:
+        async with aiohttp.ClientSession(raise_for_status=False) as session:
             payload = {
                 "model": Models.EDIT if codex is False else Models.CODE_EDIT,
                 "input": "" if text is None else text,
@@ -676,7 +681,7 @@ class Model:
         factor=3,
         base=5,
         max_tries=6,
-        on_backoff=backoff_handler,
+        on_backoff=backoff_handler_http,
     )
     async def send_moderations_request(self, text):
         # Use aiohttp to send the above request:
@@ -695,11 +700,11 @@ class Model:
 
     @backoff.on_exception(
         backoff.expo,
-        aiohttp.ClientResponseError,
+        ValueError,
         factor=3,
         base=5,
         max_tries=4,
-        on_backoff=backoff_handler,
+        on_backoff=backoff_handler_request,
     )
     async def send_summary_request(self, prompt, custom_api_key=None):
         """
@@ -718,7 +723,7 @@ class Model:
 
         tokens = self.usage_service.count_tokens(summary_request_text)
 
-        async with aiohttp.ClientSession(raise_for_status=True) as session:
+        async with aiohttp.ClientSession(raise_for_status=False) as session:
             payload = {
                 "model": Models.DAVINCI,
                 "prompt": summary_request_text,
@@ -746,11 +751,11 @@ class Model:
 
     @backoff.on_exception(
         backoff.expo,
-        aiohttp.ClientResponseError,
+        ValueError,
         factor=3,
         base=5,
         max_tries=4,
-        on_backoff=backoff_handler,
+        on_backoff=backoff_handler_request,
     )
     async def send_request(
         self,
@@ -774,20 +779,24 @@ class Model:
                 f"Prompt must be greater than {self.prompt_min_length} characters, it is currently: {len(prompt)} characters"
             )
 
+        if not max_tokens_override:
+            if model:
+                max_tokens_override = Models.get_max_tokens(model) - tokens
+
         print(f"The prompt about to be sent is {prompt}")
         print(
             f"Overrides -> temp:{temp_override}, top_p:{top_p_override} frequency:{frequency_penalty_override}, presence:{presence_penalty_override}"
         )
 
-        async with aiohttp.ClientSession(raise_for_status=True) as session:
+        async with aiohttp.ClientSession(raise_for_status=False) as session:
             payload = {
                 "model": self.model if model is None else model,
                 "prompt": prompt,
                 "stop": "" if stop is None else stop,
                 "temperature": self.temp if temp_override is None else temp_override,
                 "top_p": self.top_p if top_p_override is None else top_p_override,
-                "max_tokens": self.max_tokens - tokens
-                if not max_tokens_override
+                "max_tokens": self.max_tokens - tokens 
+                if max_tokens_override is None
                 else max_tokens_override,
                 "presence_penalty": self.presence_penalty
                 if presence_penalty_override is None
@@ -839,7 +848,7 @@ class Model:
         factor=3,
         base=5,
         max_tries=4,
-        on_backoff=backoff_handler,
+        on_backoff=backoff_handler_http,
     )
     async def send_image_request(
         self, ctx, prompt, vary=None, custom_api_key=None