0.3.2.7

19 files changed, 656 insertions, 733 deletions

diff --git a/g4f/Provider/AiChatOnline.py b/g4f/Provider/AiChatOnline.py
index 152a7d31..40f77105 100644
--- a/g4f/Provider/AiChatOnline.py
+++ b/g4f/Provider/AiChatOnline.py
@@ -12,10 +12,8 @@ class AiChatOnline(AsyncGeneratorProvider, ProviderModelMixin):
     url = "https://aichatonlineorg.erweima.ai"
     api_endpoint = "/aichatonline/api/chat/gpt"
     working = True
-    supports_gpt_35_turbo = True
     supports_gpt_4 = True
     default_model = 'gpt-4o-mini'
-    supports_message_history = False
 
     @classmethod
     async def grab_token(
diff --git a/g4f/Provider/Airforce.py b/g4f/Provider/Airforce.py
new file mode 100644
index 00000000..88896096
--- /dev/null
+++ b/g4f/Provider/Airforce.py
@@ -0,0 +1,255 @@
+from __future__ import annotations
+
+from aiohttp import ClientSession, ClientResponseError
+from urllib.parse import urlencode
+import json
+import io
+import asyncio
+
+from ..typing import AsyncResult, Messages
+from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
+from ..image import ImageResponse, is_accepted_format
+from .helper import format_prompt
+
+class Airforce(AsyncGeneratorProvider, ProviderModelMixin):
+    url = "https://api.airforce"
+    text_api_endpoint = "https://api.airforce/chat/completions"
+    image_api_endpoint = "https://api.airforce/v1/imagine2"
+    working = True
+    supports_gpt_35_turbo = True
+    supports_gpt_4 = True
+    supports_stream = True
+    supports_system_message = True
+    supports_message_history = True
+    default_model = 'llama-3-70b-chat'
+    text_models = [
+        # Open source models
+        'llama-2-13b-chat',
+        
+        'llama-3-70b-chat',
+        'llama-3-70b-chat-turbo',
+        'llama-3-70b-chat-lite',
+        
+        'llama-3-8b-chat',
+        'llama-3-8b-chat-turbo',
+        'llama-3-8b-chat-lite',
+        
+        'llama-3.1-405b-turbo',
+        'llama-3.1-70b-turbo',
+        'llama-3.1-8b-turbo',
+        
+        'LlamaGuard-2-8b',
+        'Llama-Guard-7b',
+        'Meta-Llama-Guard-3-8B',
+
+        'Mixtral-8x7B-Instruct-v0.1',
+        'Mixtral-8x22B-Instruct-v0.1',
+        'Mistral-7B-Instruct-v0.1',
+        'Mistral-7B-Instruct-v0.2',
+        'Mistral-7B-Instruct-v0.3',
+        
+        'Qwen1.5-72B-Chat',
+        'Qwen1.5-110B-Chat',
+        'Qwen2-72B-Instruct',
+        
+        'gemma-2b-it',
+        'gemma-2-9b-it',
+        'gemma-2-27b-it',
+        
+        'dbrx-instruct',
+
+        'deepseek-llm-67b-chat',
+        
+        'Nous-Hermes-2-Mixtral-8x7B-DPO',
+        'Nous-Hermes-2-Yi-34B',
+        
+        'WizardLM-2-8x22B',
+        
+        'SOLAR-10.7B-Instruct-v1.0',
+        
+        'StripedHyena-Nous-7B',      
+        
+        'sparkdesk',
+        
+        
+        # Other models
+        'chatgpt-4o-latest',
+        'gpt-4',
+        'gpt-4-turbo',
+        'gpt-4o-mini-2024-07-18',
+        'gpt-4o-mini',
+        'gpt-4o',
+        'gpt-3.5-turbo',
+        'gpt-3.5-turbo-0125',
+        'gpt-3.5-turbo-1106',
+        'gpt-3.5-turbo-16k',
+        'gpt-3.5-turbo-0613',
+        'gpt-3.5-turbo-16k-0613',
+        
+        'gemini-1.5-flash',
+        'gemini-1.5-pro',
+    ]
+    image_models = [
+        'flux',
+        'flux-realism',
+        'flux-anime',
+        'flux-3d',
+        'flux-disney',
+        'flux-pixel',
+        'any-dark',
+    ]
+    
+    models = [
+        *text_models,
+        *image_models
+    ]
+    model_aliases = {
+        # Open source models
+        "llama-2-13b": "llama-2-13b-chat",
+        
+        "llama-3-70b": "llama-3-70b-chat",
+        "llama-3-70b": "llama-3-70b-chat-turbo",
+        "llama-3-70b": "llama-3-70b-chat-lite",
+        
+        "llama-3-8b": "llama-3-8b-chat",
+        "llama-3-8b": "llama-3-8b-chat-turbo",
+        "llama-3-8b": "llama-3-8b-chat-lite",
+        
+        "llama-3.1-405b": "llama-3.1-405b-turbo",
+        "llama-3.1-70b": "llama-3.1-70b-turbo",
+        "llama-3.1-8b": "llama-3.1-8b-turbo",
+        
+        "mixtral-8x7b": "Mixtral-8x7B-Instruct-v0.1",
+        "mixtral-8x22b": "Mixtral-8x22B-Instruct-v0.1",
+        "mistral-7b": "Mistral-7B-Instruct-v0.1",
+        "mistral-7b": "Mistral-7B-Instruct-v0.2",
+        "mistral-7b": "Mistral-7B-Instruct-v0.3",
+        
+        "mixtral-8x7b-dpo": "Nous-Hermes-2-Mixtral-8x7B-DPO",
+        
+        "qwen-1-5-72b": "Qwen1.5-72B-Chat",
+        "qwen-1_5-110b": "Qwen1.5-110B-Chat",
+        "qwen-2-72b": "Qwen2-72B-Instruct",
+        
+        "gemma-2b": "gemma-2b-it",
+        "gemma-2b-9b": "gemma-2-9b-it",
+        "gemma-2b-27b": "gemma-2-27b-it",
+        
+        "deepseek": "deepseek-llm-67b-chat",
+        
+        "yi-34b": "Nous-Hermes-2-Yi-34B",
+        
+        "wizardlm-2-8x22b": "WizardLM-2-8x22B",
+        
+        "solar-10-7b": "SOLAR-10.7B-Instruct-v1.0",
+        
+        "sh-n-7b": "StripedHyena-Nous-7B",
+        
+        "sparkdesk-v1.1": "sparkdesk",
+        
+        
+        # Other models
+        "gpt-4o": "chatgpt-4o-latest",
+        "gpt-4o-mini": "gpt-4o-mini-2024-07-18",
+        
+        "gpt-3.5-turbo": "gpt-3.5-turbo-0125",
+        "gpt-3.5-turbo": "gpt-3.5-turbo-1106",
+        "gpt-3.5-turbo": "gpt-3.5-turbo-16k",
+        "gpt-3.5-turbo": "gpt-3.5-turbo-0613",
+        "gpt-3.5-turbo": "gpt-3.5-turbo-16k-0613",
+        
+        
+        "gemini-flash": "gemini-1.5-flash",
+        "gemini-pro": "gemini-1.5-pro",
+    }
+
+    @classmethod
+    async def create_async_generator(
+        cls,
+        model: str,
+        messages: Messages,
+        proxy: str = None,
+        **kwargs
+    ) -> AsyncResult:
+        model = cls.get_model(model)
+        
+        headers = {
+            "accept": "*/*",
+            "accept-language": "en-US,en;q=0.9",
+            "content-type": "application/json",
+            "origin": "https://api.airforce",
+            "sec-ch-ua": '"Chromium";v="128", "Not(A:Brand";v="24"',
+            "sec-ch-ua-mobile": "?0",
+            "sec-ch-ua-platform": '"Linux"',
+            "sec-fetch-dest": "empty",
+            "sec-fetch-mode": "cors",
+            "sec-fetch-site": "cross-site",
+            "user-agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/128.0.0.0 Safari/537.36"
+        }
+
+        
+        if model in cls.image_models:
+            async for item in cls.generate_image(model, messages, headers, proxy, **kwargs):
+                yield item
+        else:
+            async for item in cls.generate_text(model, messages, headers, proxy, **kwargs):
+                yield item
+
+    @classmethod
+    async def generate_text(cls, model: str, messages: Messages, headers: dict, proxy: str, **kwargs) -> AsyncResult:
+        async with ClientSession(headers=headers) as session:
+            data = {
+                "messages": [{"role": "user", "content": format_prompt(messages)}],
+                "model": model,
+                "temperature": kwargs.get('temperature', 1),
+                "top_p": kwargs.get('top_p', 1),
+                "stream": True
+            }
+            
+            async with session.post(cls.text_api_endpoint, json=data, proxy=proxy) as response:
+                response.raise_for_status()
+                async for line in response.content:
+                    if line:
+                        line = line.decode('utf-8').strip()
+                        if line.startswith("data: "):
+                            try:
+                                data = json.loads(line[6:])
+                                if 'choices' in data and len(data['choices']) > 0:
+                                    delta = data['choices'][0].get('delta', {})
+                                    if 'content' in delta:
+                                        yield delta['content']
+                            except json.JSONDecodeError:
+                                continue
+                        elif line == "data: [DONE]":
+                            break
+
+    @classmethod
+    async def generate_image(cls, model: str, messages: Messages, headers: dict, proxy: str, **kwargs) -> AsyncResult:
+        prompt = messages[-1]['content'] if messages else ""
+        params = {
+            "prompt": prompt,
+            "size": kwargs.get("size", "1:1"),
+            "seed": kwargs.get("seed"),
+            "model": model
+        }
+        params = {k: v for k, v in params.items() if v is not None}
+
+        try:
+            async with ClientSession(headers=headers) as session:
+                async with session.get(cls.image_api_endpoint, params=params, proxy=proxy) as response:
+                    response.raise_for_status()
+                    content = await response.read()
+                    
+                    if response.content_type.startswith('image/'):
+                        image_url = str(response.url)
+                        yield ImageResponse(image_url, prompt)
+                    else:
+                        try:
+                            text = content.decode('utf-8', errors='ignore')
+                            yield f"Error: {text}"
+                        except Exception as decode_error:
+                            yield f"Error: Unable to decode response - {str(decode_error)}"
+        except ClientResponseError as e:
+            yield f"Error: HTTP {e.status}: {e.message}"
+        except Exception as e:
+            yield f"Unexpected error: {str(e)}"
diff --git a/g4f/Provider/Bixin123.py b/g4f/Provider/Bixin123.py
index 694a2eff..39422c93 100644
--- a/g4f/Provider/Bixin123.py
+++ b/g4f/Provider/Bixin123.py
@@ -2,6 +2,7 @@ from __future__ import annotations
 
 from aiohttp import ClientSession
 import json
+import random
 from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
 from ..typing import AsyncResult, Messages
 from .helper import format_prompt
@@ -31,6 +32,10 @@ class Bixin123(AsyncGeneratorProvider, ProviderModelMixin):
             return cls.default_model
 
     @classmethod
+    def generate_fingerprint(cls) -> str:
+        return str(random.randint(100000000, 999999999))
+
+    @classmethod
     async def create_async_generator(
         cls,
         model: str,
@@ -45,7 +50,7 @@ class Bixin123(AsyncGeneratorProvider, ProviderModelMixin):
             "accept-language": "en-US,en;q=0.9",
             "cache-control": "no-cache",
             "content-type": "application/json",
-            "fingerprint": "988148794",
+            "fingerprint": cls.generate_fingerprint(),
             "origin": cls.url,
             "pragma": "no-cache",
             "priority": "u=1, i",
diff --git a/g4f/Provider/Blackbox.py b/g4f/Provider/Blackbox.py
index 9fab4a09..e607a43c 100644
--- a/g4f/Provider/Blackbox.py
+++ b/g4f/Provider/Blackbox.py
@@ -1,43 +1,40 @@
 from __future__ import annotations
 
-import uuid
-import secrets
 import re
-import base64
+import json
+import random
+import string
 from aiohttp import ClientSession
-from typing import AsyncGenerator, Optional
 
 from ..typing import AsyncResult, Messages, ImageType
-from ..image import to_data_uri, ImageResponse
+from ..image import ImageResponse, to_data_uri
 from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
 
 class Blackbox(AsyncGeneratorProvider, ProviderModelMixin):
     url = "https://www.blackbox.ai"
+    api_endpoint = "https://www.blackbox.ai/api/chat"
     working = True
+    supports_stream = True
+    supports_system_message = True
+    supports_message_history = True
+    
     default_model = 'blackbox'
     models = [
-        default_model,
-        "gemini-1.5-flash",
+        'blackbox',
+        'gemini-1.5-flash',
         "llama-3.1-8b",
         'llama-3.1-70b',
         'llama-3.1-405b',
-        'ImageGeneration',
+        'ImageGenerationLV45LJp'
     ]
-    
-    model_aliases = {
-        "gemini-flash": "gemini-1.5-flash",
-    }
-    
-    agent_mode_map = {
-        'ImageGeneration': {"mode": True, "id": "ImageGenerationLV45LJp", "name": "Image Generation"},
-    }
 
-    model_id_map = {
+    model_config = {
         "blackbox": {},
         "gemini-1.5-flash": {'mode': True, 'id': 'Gemini'},
         "llama-3.1-8b": {'mode': True, 'id': "llama-3.1-8b"},
         'llama-3.1-70b': {'mode': True, 'id': "llama-3.1-70b"},
-        'llama-3.1-405b': {'mode': True, 'id': "llama-3.1-405b"}
+        'llama-3.1-405b': {'mode': True, 'id': "llama-3.1-405b"},
+        'ImageGenerationLV45LJp': {'mode': True, 'id': "ImageGenerationLV45LJp", 'name': "Image Generation"},
     }
 
     @classmethod
@@ -50,107 +47,81 @@ class Blackbox(AsyncGeneratorProvider, ProviderModelMixin):
             return cls.default_model
 
     @classmethod
-    async def download_image_to_base64_url(cls, url: str) -> str:
-        async with ClientSession() as session:
-            async with session.get(url) as response:
-                image_data = await response.read()
-                base64_data = base64.b64encode(image_data).decode('utf-8')
-                mime_type = response.headers.get('Content-Type', 'image/jpeg')
-                return f"data:{mime_type};base64,{base64_data}"
-
-    @classmethod
     async def create_async_generator(
         cls,
         model: str,
         messages: Messages,
-        proxy: Optional[str] = None,
-        image: Optional[ImageType] = None,
-        image_name: Optional[str] = None,
+        proxy: str = None,
+        image: ImageType = None,
+        image_name: str = None,
         **kwargs
-    ) -> AsyncGenerator[AsyncResult, None]:
-        if image is not None:
-            messages[-1]["data"] = {
-                "fileText": image_name,
-                "imageBase64": to_data_uri(image),
-                "title": str(uuid.uuid4())
-            }
-
+    ) -> AsyncResult:
+        model = cls.get_model(model)
+        
         headers = {
-            "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36",
-            "Accept": "*/*",
-            "Accept-Language": "en-US,en;q=0.5",
-            "Accept-Encoding": "gzip, deflate, br",
-            "Referer": cls.url,
-            "Content-Type": "application/json",
-            "Origin": cls.url,
-            "DNT": "1",
-            "Sec-GPC": "1",
-            "Alt-Used": "www.blackbox.ai",
-            "Connection": "keep-alive",
+            "accept": "*/*",
+            "accept-language": "en-US,en;q=0.9",
+            "cache-control": "no-cache",
+            "content-type": "application/json",
+            "origin": cls.url,
+            "pragma": "no-cache",
+            "referer": f"{cls.url}/",
+            "sec-ch-ua": '"Not;A=Brand";v="24", "Chromium";v="128"',
+            "sec-ch-ua-mobile": "?0",
+            "sec-ch-ua-platform": '"Linux"',
+            "sec-fetch-dest": "empty",
+            "sec-fetch-mode": "cors",
+            "sec-fetch-site": "same-origin",
+            "user-agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/128.0.0.0 Safari/537.36"
         }
-
+        
         async with ClientSession(headers=headers) as session:
-            random_id = secrets.token_hex(16)
-            random_user_id = str(uuid.uuid4())
-            
-            model = cls.get_model(model)  # Resolve the model alias
+            if image is not None:
+                messages[-1]["data"] = {
+                    "fileText": image_name,
+                    "imageBase64": to_data_uri(image)
+                }
             
+            random_id = ''.join(random.choices(string.ascii_letters + string.digits, k=7))
+
             data = {
                 "messages": messages,
                 "id": random_id,
-                "userId": random_user_id,
+                "previewToken": None,
+                "userId": None,
                 "codeModelMode": True,
-                "agentMode": cls.agent_mode_map.get(model, {}),
+                "agentMode": {},
                 "trendingAgentMode": {},
                 "isMicMode": False,
+                "maxTokens": None,
                 "isChromeExt": False,
-                "playgroundMode": False,
-                "webSearchMode": False,
-                "userSystemPrompt": "",
                 "githubToken": None,
-                "trendingAgentModel": cls.model_id_map.get(model, {}),
-                "maxTokens": None
+                "clickedAnswer2": False,
+                "clickedAnswer3": False,
+                "clickedForceWebSearch": False,
+                "visitFromDelta": False,
+                "mobileClient": False
             }
 
-            async with session.post(
-                f"{cls.url}/api/chat", json=data, proxy=proxy
-            ) as response:
+            if model == 'ImageGenerationLV45LJp':
+                data["agentMode"] = cls.model_config[model]
+            else:
+                data["trendingAgentMode"] = cls.model_config[model]
+            
+            async with session.post(cls.api_endpoint, json=data, proxy=proxy) as response:
                 response.raise_for_status()
-                full_response = ""
-                buffer = ""
-                image_base64_url = None
-                async for chunk in response.content.iter_any():
-                    if chunk:
-                        decoded_chunk = chunk.decode()
-                        cleaned_chunk = re.sub(r'\$@\$.+?\$@\$|\$@\$', '', decoded_chunk)
-                        
-                        buffer += cleaned_chunk
-                        
-                        # Check if there's a complete image line in the buffer
-                        image_match = re.search(r'!\[Generated Image\]\((https?://[^\s\)]+)\)', buffer)
-                        if image_match:
-                            image_url = image_match.group(1)
-                            # Download the image and convert to base64 URL
-                            image_base64_url = await cls.download_image_to_base64_url(image_url)
-                            
-                            # Remove the image line from the buffer
-                            buffer = re.sub(r'!\[Generated Image\]\(https?://[^\s\)]+\)', '', buffer)
-                        
-                        # Send text line by line
-                        lines = buffer.split('\n')
-                        for line in lines[:-1]:
-                            if line.strip():
-                                full_response += line + '\n'
-                                yield line + '\n'
-                        buffer = lines[-1]  # Keep the last incomplete line in the buffer
-
-                # Send the remaining buffer if it's not empty
-                if buffer.strip():
-                    full_response += buffer
-                    yield buffer
-
-                # If an image was found, send it as ImageResponse
-                if image_base64_url:
-                    alt_text = "Generated Image"
-                    image_response = ImageResponse(image_base64_url, alt=alt_text)
-                    yield image_response
+                if model == 'ImageGenerationLV45LJp':
+                    response_text = await response.text()
+                    url_match = re.search(r'https://storage\.googleapis\.com/[^\s\)]+', response_text)
+                    if url_match:
+                        image_url = url_match.group(0)
+                        yield ImageResponse(image_url, alt=messages[-1]['content'])
+                    else:
+                        raise Exception("Image URL not found in the response")
+                else:
+                    async for chunk in response.content:
+                        if chunk:
+                            decoded_chunk = chunk.decode()
+                            if decoded_chunk.startswith('$@$v=undefined-rv1$@$'):
+                                decoded_chunk = decoded_chunk[len('$@$v=undefined-rv1$@$'):]
+                            yield decoded_chunk
diff --git a/g4f/Provider/FluxAirforce.py b/g4f/Provider/FluxAirforce.py
deleted file mode 100644
index fe003a61..00000000
--- a/g4f/Provider/FluxAirforce.py
+++ /dev/null
@@ -1,82 +0,0 @@
-from __future__ import annotations
-
-from aiohttp import ClientSession, ClientResponseError
-from urllib.parse import urlencode
-import io
-
-from ..typing import AsyncResult, Messages
-from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
-from ..image import ImageResponse, is_accepted_format
-
-class FluxAirforce(AsyncGeneratorProvider, ProviderModelMixin):
-    url = "https://flux.api.airforce/"
-    api_endpoint = "https://api.airforce/v1/imagine2"
-    working = True
-    default_model = 'flux-realism'
-    models = [
-        'flux',
-        'flux-realism',
-        'flux-anime',
-        'flux-3d',
-        'flux-disney'
-    ]
-
-    @classmethod
-    async def create_async_generator(
-        cls,
-        model: str,
-        messages: Messages,
-        proxy: str = None,
-        **kwargs
-    ) -> AsyncResult:
-        headers = {
-            "accept": "*/*",
-            "accept-language": "en-US,en;q=0.9",
-            "origin": "https://flux.api.airforce",
-            "priority": "u=1, i",
-            "referer": "https://flux.api.airforce/",
-            "sec-ch-ua": '"Chromium";v="127", "Not)A;Brand";v="99"',
-            "sec-ch-ua-mobile": "?0",
-            "sec-ch-ua-platform": '"Linux"',
-            "sec-fetch-dest": "empty",
-            "sec-fetch-mode": "cors",
-            "sec-fetch-site": "same-site",
-            "user-agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/127.0.0.0 Safari/537.36"
-        }
-
-        prompt = messages[-1]['content'] if messages else ""
-
-        params = {
-            "prompt": prompt,
-            "size": kwargs.get("size", "1:1"),
-            "seed": kwargs.get("seed"),
-            "model": model
-        }
-
-        params = {k: v for k, v in params.items() if v is not None}
-
-        try:
-            async with ClientSession(headers=headers) as session:
-                async with session.get(f"{cls.api_endpoint}", params=params, proxy=proxy) as response:
-                    response.raise_for_status()
-                    
-                    content = await response.read()
-                    
-                    if response.content_type.startswith('image/'):
-                        image_url = str(response.url)
-                        yield ImageResponse(image_url, prompt)
-                    else:
-                        try:
-                            text = content.decode('utf-8', errors='ignore')
-                            yield f"Error: {text}"
-                        except Exception as decode_error:
-                            yield f"Error: Unable to decode response - {str(decode_error)}"
-
-        except ClientResponseError as e:
-            yield f"Error: HTTP {e.status}: {e.message}"
-        except Exception as e:
-            yield f"Unexpected error: {str(e)}"
-
-        finally:
-            if not session.closed:
-                await session.close()
diff --git a/g4f/Provider/HuggingChat.py b/g4f/Provider/HuggingChat.py
index 76c76a35..06216ade 100644
--- a/g4f/Provider/HuggingChat.py
+++ b/g4f/Provider/HuggingChat.py
@@ -12,24 +12,21 @@ class HuggingChat(AbstractProvider, ProviderModelMixin):
     working = True
     supports_stream = True
     default_model = "meta-llama/Meta-Llama-3.1-70B-Instruct"
+    
     models = [
         'meta-llama/Meta-Llama-3.1-70B-Instruct',
-        'meta-llama/Meta-Llama-3.1-405B-Instruct-FP8',
-        'CohereForAI/c4ai-command-r-plus',
+        'CohereForAI/c4ai-command-r-plus-08-2024',
         'mistralai/Mixtral-8x7B-Instruct-v0.1',
         'NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO',
-        '01-ai/Yi-1.5-34B-Chat',
         'mistralai/Mistral-7B-Instruct-v0.3',
         'microsoft/Phi-3-mini-4k-instruct',
     ]
     
     model_aliases = {
         "llama-3.1-70b": "meta-llama/Meta-Llama-3.1-70B-Instruct",
-        "llama-3.1-405b": "meta-llama/Meta-Llama-3.1-405B-Instruct-FP8",
-        "command-r-plus": "CohereForAI/c4ai-command-r-plus",
+        "command-r-plus": "CohereForAI/c4ai-command-r-plus-08-2024",
         "mixtral-8x7b": "mistralai/Mixtral-8x7B-Instruct-v0.1",
-        "mixtral-8x7b": "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
-        "yi-1.5-34b": "01-ai/Yi-1.5-34B-Chat",
+        "mixtral-8x7b-dpo": "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
         "mistral-7b": "mistralai/Mistral-7B-Instruct-v0.3",
         "phi-3-mini-4k": "microsoft/Phi-3-mini-4k-instruct",
     }
@@ -80,7 +77,7 @@ class HuggingChat(AbstractProvider, ProviderModelMixin):
             response = session.post('https://huggingface.co/chat/conversation', json=json_data)
             conversationId = response.json()['conversationId']
 
-            response = session.get(f'https://huggingface.co/chat/conversation/{conversationId}/__data.json?x-sveltekit-invalidated=01',)
+            response = session.get(f'https://huggingface.co/chat/conversation/{conversationId}/__data.json?x-sveltekit-invalidated=11',)
 
             data: list = (response.json())["nodes"][1]["data"]
             keys: list[int] = data[data[0]["messages"]]
diff --git a/g4f/Provider/HuggingFace.py b/g4f/Provider/HuggingFace.py
index 74957862..586e5f5f 100644
--- a/g4f/Provider/HuggingFace.py
+++ b/g4f/Provider/HuggingFace.py
@@ -9,33 +9,16 @@ from .helper import get_connector
 from ..errors import RateLimitError, ModelNotFoundError
 from ..requests.raise_for_status import raise_for_status
 
+from .HuggingChat import HuggingChat
+
 class HuggingFace(AsyncGeneratorProvider, ProviderModelMixin):
     url = "https://huggingface.co/chat"
     working = True
     needs_auth = True
     supports_message_history = True
-    default_model = "meta-llama/Meta-Llama-3.1-70B-Instruct"
-    models = [
-        'meta-llama/Meta-Llama-3.1-70B-Instruct',
-        'meta-llama/Meta-Llama-3.1-405B-Instruct-FP8',
-        'CohereForAI/c4ai-command-r-plus',
-        'mistralai/Mixtral-8x7B-Instruct-v0.1',
-        'NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO',
-        '01-ai/Yi-1.5-34B-Chat',
-        'mistralai/Mistral-7B-Instruct-v0.3',
-        'microsoft/Phi-3-mini-4k-instruct',
-    ]
-    
-    model_aliases = {
-        "llama-3.1-70b": "meta-llama/Meta-Llama-3.1-70B-Instruct",
-        "llama-3.1-405b": "meta-llama/Meta-Llama-3.1-405B-Instruct-FP8",
-        "command-r-plus": "CohereForAI/c4ai-command-r-plus",
-        "mixtral-8x7b": "mistralai/Mixtral-8x7B-Instruct-v0.1",
-        "mixtral-8x7b": "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
-        "yi-1.5-34b": "01-ai/Yi-1.5-34B-Chat",
-        "mistral-7b": "mistralai/Mistral-7B-Instruct-v0.3",
-        "phi-3-mini-4k": "microsoft/Phi-3-mini-4k-instruct",
-    }
+    default_model = HuggingChat.default_model
+    models = HuggingChat.models
+    model_aliases = HuggingChat.model_aliases
 
     @classmethod
     def get_model(cls, model: str) -> str:
diff --git a/g4f/Provider/Koala.py b/g4f/Provider/Koala.py
index 0e810083..14e533df 100644
--- a/g4f/Provider/Koala.py
+++ b/g4f/Provider/Koala.py
@@ -10,7 +10,8 @@ from .helper import get_random_string, get_connector
 from ..requests import raise_for_status
 
 class Koala(AsyncGeneratorProvider, ProviderModelMixin):
-    url = "https://koala.sh"
+    url = "https://koala.sh/chat"
+    api_endpoint = "https://koala.sh/api/gpt/"
     working = True
     supports_message_history = True
     supports_gpt_4 = True
@@ -26,17 +27,17 @@ class Koala(AsyncGeneratorProvider, ProviderModelMixin):
         **kwargs: Any
     ) -> AsyncGenerator[Dict[str, Union[str, int, float, List[Dict[str, Any]], None]], None]:
         if not model:
-            model = "gpt-3.5-turbo"
+            model = "gpt-4o-mini"
 
         headers = {
             "User-Agent": "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:122.0) Gecko/20100101 Firefox/122.0",
             "Accept": "text/event-stream",
             "Accept-Language": "de,en-US;q=0.7,en;q=0.3",
             "Accept-Encoding": "gzip, deflate, br",
-            "Referer": f"{cls.url}/chat",
+            "Referer": f"{cls.url}",
             "Flag-Real-Time-Data": "false",
             "Visitor-ID": get_random_string(20),
-            "Origin": cls.url,
+            "Origin": "https://koala.sh",
             "Alt-Used": "koala.sh",
             "Sec-Fetch-Dest": "empty",
             "Sec-Fetch-Mode": "cors",
@@ -67,7 +68,7 @@ class Koala(AsyncGeneratorProvider, ProviderModelMixin):
                 "model": model,
             }
 
-            async with session.post(f"{cls.url}/api/gpt/", json=data, proxy=proxy) as response:
+            async with session.post(f"{cls.api_endpoint}", json=data, proxy=proxy) as response:
                 await raise_for_status(response)
                 async for chunk in cls._parse_event_stream(response):
                     yield chunk
diff --git a/g4f/Provider/Llama.py b/g4f/Provider/Llama.py
deleted file mode 100644
index 235c0994..00000000
--- a/g4f/Provider/Llama.py
+++ /dev/null
@@ -1,91 +0,0 @@
-from __future__ import annotations
-
-from aiohttp import ClientSession
-
-from ..typing import AsyncResult, Messages
-from ..requests.raise_for_status import raise_for_status
-from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
-
-
-class Llama(AsyncGeneratorProvider, ProviderModelMixin):
-    url = "https://www.llama2.ai"
-    working = False
-    supports_message_history = True
-    default_model = "meta/meta-llama-3-70b-instruct"
-    models = [
-        "meta/llama-2-7b-chat",
-        "meta/llama-2-13b-chat",
-        "meta/llama-2-70b-chat",
-        "meta/meta-llama-3-8b-instruct",
-        "meta/meta-llama-3-70b-instruct",
-    ]
-    model_aliases = {
-        "meta-llama/Meta-Llama-3-8B-Instruct": "meta/meta-llama-3-8b-instruct",
-        "meta-llama/Meta-Llama-3-70B-Instruct": "meta/meta-llama-3-70b-instruct",
-        "meta-llama/Llama-2-7b-chat-hf": "meta/llama-2-7b-chat",
-        "meta-llama/Llama-2-13b-chat-hf": "meta/llama-2-13b-chat",
-        "meta-llama/Llama-2-70b-chat-hf": "meta/llama-2-70b-chat",
-    }
-
-    @classmethod
-    async def create_async_generator(
-        cls,
-        model: str,
-        messages: Messages,
-        proxy: str = None,
-        system_message: str = "You are a helpful assistant.",
-        temperature: float = 0.75,
-        top_p: float = 0.9,
-        max_tokens: int = 8000,
-        **kwargs
-    ) -> AsyncResult:
-        headers = {
-            "User-Agent": "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/118.0",
-            "Accept": "*/*",
-            "Accept-Language": "de,en-US;q=0.7,en;q=0.3",
-            "Accept-Encoding": "gzip, deflate, br",
-            "Referer": f"{cls.url}/",
-            "Content-Type": "text/plain;charset=UTF-8",
-            "Origin": cls.url,
-            "Connection": "keep-alive",
-            "Sec-Fetch-Dest": "empty",
-            "Sec-Fetch-Mode": "cors",
-            "Sec-Fetch-Site": "same-origin",
-            "Pragma": "no-cache",
-            "Cache-Control": "no-cache",
-            "TE": "trailers"
-        }
-        async with ClientSession(headers=headers) as session:
-            system_messages = [message["content"] for message in messages if message["role"] == "system"]
-            if system_messages:
-                system_message = "\n".join(system_messages)
-                messages = [message for message in messages if message["role"] != "system"] 
-            prompt = format_prompt(messages)
-            data = {
-                "prompt": prompt,
-                "model": cls.get_model(model),
-                "systemPrompt": system_message,
-                "temperature": temperature,
-                "topP": top_p,
-                "maxTokens": max_tokens,
-                "image": None
-            }
-            started = False
-            async with session.post(f"{cls.url}/api", json=data, proxy=proxy) as response:
-                await raise_for_status(response)
-                async for chunk in response.content.iter_any():
-                    if not chunk:
-                        continue
-                    if not started:
-                        chunk = chunk.lstrip()
-                        started = True
-                    yield chunk.decode(errors="ignore")
-            
-def format_prompt(messages: Messages):
-    messages = [
-        f"[INST] {message['content']} [/INST]"
-        if message["role"] == "user"
-        else message["content"]
-        for message in messages
-    ]
-    return "\n".join(messages) + "\n"
diff --git a/g4f/Provider/Nexra.py b/g4f/Provider/Nexra.py
index e2c3e197..b2b83837 100644
--- a/g4f/Provider/Nexra.py
+++ b/g4f/Provider/Nexra.py
@@ -1,40 +1,32 @@
 from __future__ import annotations
-
 import json
-import base64
 from aiohttp import ClientSession
-from typing import AsyncGenerator
 
 from ..typing import AsyncResult, Messages
 from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
-from ..image import ImageResponse
 from .helper import format_prompt
+from ..image import ImageResponse
 
 class Nexra(AsyncGeneratorProvider, ProviderModelMixin):
     url = "https://nexra.aryahcr.cc"
-    api_endpoint_text = "https://nexra.aryahcr.cc/api/chat/gpt"
-    api_endpoint_image = "https://nexra.aryahcr.cc/api/image/complements"
+    chat_api_endpoint = "https://nexra.aryahcr.cc/api/chat/gpt"
+    image_api_endpoint = "https://nexra.aryahcr.cc/api/image/complements"
     working = True
     supports_gpt_35_turbo = True
     supports_gpt_4 = True
-    supports_stream = True
     supports_system_message = True
     supports_message_history = True
     
     default_model = 'gpt-3.5-turbo'
-    models = [
-        # Text models
+    text_models = [
         'gpt-4', 'gpt-4-0613', 'gpt-4-32k', 'gpt-4-0314', 'gpt-4-32k-0314',
         'gpt-3.5-turbo', 'gpt-3.5-turbo-16k', 'gpt-3.5-turbo-0613', 'gpt-3.5-turbo-16k-0613', 'gpt-3.5-turbo-0301',
         'gpt-3', 'text-davinci-003', 'text-davinci-002', 'code-davinci-002',
         'text-curie-001', 'text-babbage-001', 'text-ada-001',
         'davinci', 'curie', 'babbage', 'ada', 'babbage-002', 'davinci-002',
-        # Image models
-        'dalle', 'dalle-mini', 'emi'
     ]
-    
-    image_models = {"dalle", "dalle-mini", "emi"}
-    text_models = set(models) - image_models
+    image_models = ['dalle', 'dalle2', 'dalle-mini', 'emi']
+    models = [*text_models, *image_models]
     
     model_aliases = {
         "gpt-4": "gpt-4-0613",
@@ -60,11 +52,13 @@ class Nexra(AsyncGeneratorProvider, ProviderModelMixin):
         "gpt-3": "ada",
         "gpt-3": "babbage-002",
         "gpt-3": "davinci-002",
+        
+        "dalle-2": "dalle2",
     }
-
+    
     @classmethod
     def get_model(cls, model: str) -> str:
-        if model in cls.models:
+        if model in cls.text_models or model in cls.image_models:
             return model
         elif model in cls.model_aliases:
             return cls.model_aliases[model]
@@ -78,104 +72,45 @@ class Nexra(AsyncGeneratorProvider, ProviderModelMixin):
         messages: Messages,
         proxy: str = None,
         **kwargs
-    ) -> AsyncGenerator[str | ImageResponse, None]:
+    ) -> AsyncResult:
         model = cls.get_model(model)
         
-        if model in cls.image_models:
-            async for result in cls.create_image_async_generator(model, messages, proxy, **kwargs):
-                yield result
-        else:
-            async for result in cls.create_text_async_generator(model, messages, proxy, **kwargs):
-                yield result
-
-    @classmethod
-    async def create_text_async_generator(
-        cls,
-        model: str,
-        messages: Messages,
-        proxy: str = None,
-        **kwargs
-    ) -> AsyncGenerator[str, None]:
         headers = {
             "Content-Type": "application/json",
         }
+        
         async with ClientSession(headers=headers) as session:
-            data = {
-                "messages": messages,
-                "prompt": format_prompt(messages),
-                "model": model,
-                "markdown": False,
-                "stream": False,
-            }
-            async with session.post(cls.api_endpoint_text, json=data, proxy=proxy) as response:
-                response.raise_for_status()
-                result = await response.text()
-                json_result = json.loads(result)
-                yield json_result["gpt"]
-
-    @classmethod
-    async def create_image_async_generator(
-        cls,
-        model: str,
-        messages: Messages,
-        proxy: str = None,
-        **kwargs
-    ) -> AsyncGenerator[ImageResponse | str, None]:
-        headers = {
-            "Content-Type": "application/json"
-        }
-
-        prompt = messages[-1]['content'] if messages else ""
-
-        data = {
-            "prompt": prompt,
-            "model": model
-        }
-
-        async def process_response(response_text: str) -> ImageResponse | None:
-            json_start = response_text.find('{')
-            if json_start != -1:
-                json_data = response_text[json_start:]
-                try:
-                    response_data = json.loads(json_data)
-                    image_data = response_data.get('images', [])[0]
+            if model in cls.image_models:
+                # Image generation
+                prompt = messages[-1]['content'] if messages else ""
+                data = {
+                    "prompt": prompt,
+                    "model": model,
+                    "response": "url"
+                }
+                async with session.post(cls.image_api_endpoint, json=data, proxy=proxy) as response:
+                    response.raise_for_status()
+                    result = await response.text()
+                    result_json = json.loads(result.strip('_'))
+                    image_url = result_json['images'][0] if result_json['images'] else None
                     
-                    if image_data.startswith('data:image/'):
-                        return ImageResponse([image_data], "Generated image")
+                    if image_url:
+                        yield ImageResponse(images=image_url, alt=prompt)
+            else:
+                # Text completion
+                data = {
+                    "messages": messages,
+                    "prompt": format_prompt(messages),
+                    "model": model,
+                    "markdown": False
+                }
+                async with session.post(cls.chat_api_endpoint, json=data, proxy=proxy) as response:
+                    response.raise_for_status()
+                    result = await response.text()
                     
                     try:
-                        base64.b64decode(image_data)
-                        data_uri = f"data:image/jpeg;base64,{image_data}"
-                        return ImageResponse([data_uri], "Generated image")
-                    except:
-                        print("Invalid base64 data")
-                        return None
-                except json.JSONDecodeError:
-                    print("Failed to parse JSON.")
-            else:
-                print("No JSON data found in the response.")
-            return None
-
-        async with ClientSession(headers=headers) as session:
-            async with session.post(cls.api_endpoint_image, json=data, proxy=proxy) as response:
-                response.raise_for_status()
-                response_text = await response.text()
-                
-                image_response = await process_response(response_text)
-                if image_response:
-                    yield image_response
-                else:
-                    yield "Failed to process image data."
-
-    @classmethod
-    async def create_async(
-        cls,
-        model: str,
-        messages: Messages,
-        proxy: str = None,
-        **kwargs
-    ) -> str:
-        async for response in cls.create_async_generator(model, messages, proxy, **kwargs):
-            if isinstance(response, ImageResponse):
-                return response.images[0]
-            return response
+                        json_response = json.loads(result)
+                        gpt_response = json_response.get('gpt', '')
+                        yield gpt_response
+                    except json.JSONDecodeError:
+                        yield result
diff --git a/g4f/Provider/PerplexityLabs.py b/g4f/Provider/PerplexityLabs.py
index 3656a39b..ecb51f9b 100644
--- a/g4f/Provider/PerplexityLabs.py
+++ b/g4f/Provider/PerplexityLabs.py
@@ -13,7 +13,7 @@ WS_URL = "wss://www.perplexity.ai/socket.io/"
 class PerplexityLabs(AsyncGeneratorProvider, ProviderModelMixin):
     url = "https://labs.perplexity.ai"
     working = True
-    default_model = "llama-3.1-8b-instruct"
+    default_model = "llama-3.1-70b-instruct"
     models = [
         "llama-3.1-sonar-large-128k-online",
         "llama-3.1-sonar-small-128k-online",
@@ -22,6 +22,15 @@ class PerplexityLabs(AsyncGeneratorProvider, ProviderModelMixin):
         "llama-3.1-8b-instruct",
         "llama-3.1-70b-instruct",
     ]
+    
+    model_aliases = {
+        "llama-3.1-8b": "llama-3.1-sonar-large-128k-online",
+        "llama-3.1-8b": "sonar-small-128k-online",
+        "llama-3.1-8b": "llama-3.1-sonar-large-128k-chat",
+        "llama-3.1-8b": "llama-3.1-sonar-small-128k-chat",
+        "llama-3.1-8b": "llama-3.1-8b-instruct",
+        "llama-3.1-70b": "llama-3.1-70b-instruct",
+    }
 
     @classmethod
     async def create_async_generator(
diff --git a/g4f/Provider/Prodia.py b/g4f/Provider/Prodia.py
new file mode 100644
index 00000000..dd87a34c
--- /dev/null
+++ b/g4f/Provider/Prodia.py
@@ -0,0 +1,149 @@
+from __future__ import annotations
+
+from aiohttp import ClientSession
+import time
+import asyncio
+
+from ..typing import AsyncResult, Messages
+from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
+from ..image import ImageResponse
+
+class Prodia(AsyncGeneratorProvider, ProviderModelMixin):
+    url = "https://app.prodia.com"
+    api_endpoint = "https://api.prodia.com/generate"
+    working = True
+    
+    default_model = 'absolutereality_v181.safetensors [3d9d4d2b]'
+    models = [
+        '3Guofeng3_v34.safetensors [50f420de]',
+        'absolutereality_V16.safetensors [37db0fc3]',
+        'absolutereality_v181.safetensors [3d9d4d2b]',
+        'amIReal_V41.safetensors [0a8a2e61]',
+        'analog-diffusion-1.0.ckpt [9ca13f02]',
+        'aniverse_v30.safetensors [579e6f85]',
+        'anythingv3_0-pruned.ckpt [2700c435]',
+        'anything-v4.5-pruned.ckpt [65745d25]',
+        'anythingV5_PrtRE.safetensors [893e49b9]',
+        'AOM3A3_orangemixs.safetensors [9600da17]',
+        'blazing_drive_v10g.safetensors [ca1c1eab]',
+        'breakdomain_I2428.safetensors [43cc7d2f]',
+        'breakdomain_M2150.safetensors [15f7afca]',
+        'cetusMix_Version35.safetensors [de2f2560]',
+        'childrensStories_v13D.safetensors [9dfaabcb]',
+        'childrensStories_v1SemiReal.safetensors [a1c56dbb]',
+        'childrensStories_v1ToonAnime.safetensors [2ec7b88b]',
+        'Counterfeit_v30.safetensors [9e2a8f19]',
+        'cuteyukimixAdorable_midchapter3.safetensors [04bdffe6]',
+        'cyberrealistic_v33.safetensors [82b0d085]',
+        'dalcefo_v4.safetensors [425952fe]',
+        'deliberate_v2.safetensors [10ec4b29]',
+        'deliberate_v3.safetensors [afd9d2d4]',
+        'dreamlike-anime-1.0.safetensors [4520e090]',
+        'dreamlike-diffusion-1.0.safetensors [5c9fd6e0]',
+        'dreamlike-photoreal-2.0.safetensors [fdcf65e7]',
+        'dreamshaper_6BakedVae.safetensors [114c8abb]',
+        'dreamshaper_7.safetensors [5cf5ae06]',
+        'dreamshaper_8.safetensors [9d40847d]',
+        'edgeOfRealism_eorV20.safetensors [3ed5de15]',
+        'EimisAnimeDiffusion_V1.ckpt [4f828a15]',
+        'elldreths-vivid-mix.safetensors [342d9d26]',
+        'epicphotogasm_xPlusPlus.safetensors [1a8f6d35]',
+        'epicrealism_naturalSinRC1VAE.safetensors [90a4c676]',
+        'epicrealism_pureEvolutionV3.safetensors [42c8440c]',
+        'ICantBelieveItsNotPhotography_seco.safetensors [4e7a3dfd]',
+        'indigoFurryMix_v75Hybrid.safetensors [91208cbb]',
+        'juggernaut_aftermath.safetensors [5e20c455]',
+        'lofi_v4.safetensors [ccc204d6]',
+        'lyriel_v16.safetensors [68fceea2]',
+        'majicmixRealistic_v4.safetensors [29d0de58]',
+        'mechamix_v10.safetensors [ee685731]',
+        'meinamix_meinaV9.safetensors [2ec66ab0]',
+        'meinamix_meinaV11.safetensors [b56ce717]',
+        'neverendingDream_v122.safetensors [f964ceeb]',
+        'openjourney_V4.ckpt [ca2f377f]',
+        'pastelMixStylizedAnime_pruned_fp16.safetensors [793a26e8]',
+        'portraitplus_V1.0.safetensors [1400e684]',
+        'protogenx34.safetensors [5896f8d5]',
+        'Realistic_Vision_V1.4-pruned-fp16.safetensors [8d21810b]',
+        'Realistic_Vision_V2.0.safetensors [79587710]',
+        'Realistic_Vision_V4.0.safetensors [29a7afaa]',
+        'Realistic_Vision_V5.0.safetensors [614d1063]',
+        'Realistic_Vision_V5.1.safetensors [a0f13c83]',
+        'redshift_diffusion-V10.safetensors [1400e684]',
+        'revAnimated_v122.safetensors [3f4fefd9]',
+        'rundiffusionFX25D_v10.safetensors [cd12b0ee]',
+        'rundiffusionFX_v10.safetensors [cd4e694d]',
+        'sdv1_4.ckpt [7460a6fa]',
+        'v1-5-pruned-emaonly.safetensors [d7049739]',
+        'v1-5-inpainting.safetensors [21c7ab71]',
+        'shoninsBeautiful_v10.safetensors [25d8c546]',
+        'theallys-mix-ii-churned.safetensors [5d9225a4]',
+        'timeless-1.0.ckpt [7c4971d4]',
+        'toonyou_beta6.safetensors [980f6b15]',
+    ]
+
+    @classmethod
+    def get_model(cls, model: str) -> str:
+        if model in cls.models:
+            return model
+        elif model in cls.model_aliases:
+            return cls.model_aliases[model]
+        else:
+            return cls.default_model
+
+    @classmethod
+    async def create_async_generator(
+        cls,
+        model: str,
+        messages: Messages,
+        proxy: str = None,
+        **kwargs
+    ) -> AsyncResult:
+        model = cls.get_model(model)
+        
+        headers = {
+            "accept": "*/*",
+            "accept-language": "en-US,en;q=0.9",
+            "origin": cls.url,
+            "referer": f"{cls.url}/",
+            "user-agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/128.0.0.0 Safari/537.36"
+        }
+        
+        async with ClientSession(headers=headers) as session:
+            prompt = messages[-1]['content'] if messages else ""
+            
+            params = {
+                "new": "true",
+                "prompt": prompt,
+                "model": model,
+                "negative_prompt": kwargs.get("negative_prompt", ""),
+                "steps": kwargs.get("steps", 20),
+                "cfg": kwargs.get("cfg", 7),
+                "seed": kwargs.get("seed", int(time.time())),
+                "sampler": kwargs.get("sampler", "DPM++ 2M Karras"),
+                "aspect_ratio": kwargs.get("aspect_ratio", "square")
+            }
+            
+            async with session.get(cls.api_endpoint, params=params, proxy=proxy) as response:
+                response.raise_for_status()
+                job_data = await response.json()
+                job_id = job_data["job"]
+                
+                image_url = await cls._poll_job(session, job_id, proxy)
+                yield ImageResponse(image_url, alt=prompt)
+
+    @classmethod
+    async def _poll_job(cls, session: ClientSession, job_id: str, proxy: str, max_attempts: int = 30, delay: int = 2) -> str:
+        for _ in range(max_attempts):
+            async with session.get(f"https://api.prodia.com/job/{job_id}", proxy=proxy) as response:
+                response.raise_for_status()
+                job_status = await response.json()
+
+                if job_status["status"] == "succeeded":
+                    return f"https://images.prodia.xyz/{job_id}.png"
+                elif job_status["status"] == "failed":
+                    raise Exception("Image generation failed")
+
+            await asyncio.sleep(delay)
+
+        raise Exception("Timeout waiting for image generation")
diff --git a/g4f/Provider/ReplicateHome.py b/g4f/Provider/ReplicateHome.py
index c4e52ad6..7f443a7d 100644
--- a/g4f/Provider/ReplicateHome.py
+++ b/g4f/Provider/ReplicateHome.py
@@ -1,66 +1,60 @@
 from __future__ import annotations
-from typing import Generator, Optional, Dict, Any, Union, List
-import random
+
+import json
 import asyncio
-import base64
+from aiohttp import ClientSession, ContentTypeError
 
-from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
 from ..typing import AsyncResult, Messages
-from ..requests import StreamSession, raise_for_status
-from ..errors import ResponseError
+from .base_provider import AsyncGeneratorProvider, ProviderModelMixin
+from .helper import format_prompt
 from ..image import ImageResponse
 
 class ReplicateHome(AsyncGeneratorProvider, ProviderModelMixin):
     url = "https://replicate.com"
-    parent = "Replicate"
+    api_endpoint = "https://homepage.replicate.com/api/prediction"
     working = True
+    supports_stream = True
+    supports_system_message = True
+    supports_message_history = True
+    
     default_model = 'meta/meta-llama-3-70b-instruct'
-    models = [
-        # Models for image generation
-        'stability-ai/stable-diffusion-3',
-        'bytedance/sdxl-lightning-4step',
-        'playgroundai/playground-v2.5-1024px-aesthetic',
-        
-        # Models for image generation
+    
+    text_models = [
         'meta/meta-llama-3-70b-instruct',
         'mistralai/mixtral-8x7b-instruct-v0.1',
         'google-deepmind/gemma-2b-it',
+        'yorickvp/llava-13b',
     ]
 
-    versions = {
-        # Model versions for generating images
-        'stability-ai/stable-diffusion-3': [
-            "527d2a6296facb8e47ba1eaf17f142c240c19a30894f437feee9b91cc29d8e4f"
-        ],
-        'bytedance/sdxl-lightning-4step': [
-            "5f24084160c9089501c1b3545d9be3c27883ae2239b6f412990e82d4a6210f8f"
-        ],
-        'playgroundai/playground-v2.5-1024px-aesthetic': [
-            "a45f82a1382bed5c7aeb861dac7c7d191b0fdf74d8d57c4a0e6ed7d4d0bf7d24"
-        ],
-        
-        # Model versions for text generation
-        'meta/meta-llama-3-70b-instruct': [
-            "dp-cf04fe09351e25db628e8b6181276547"
-        ],
-        'mistralai/mixtral-8x7b-instruct-v0.1': [
-            "dp-89e00f489d498885048e94f9809fbc76"
-        ],
-        'google-deepmind/gemma-2b-it': [
-            "dff94eaf770e1fc211e425a50b51baa8e4cac6c39ef074681f9e39d778773626"
-        ]
-    }
-
-    image_models = {"stability-ai/stable-diffusion-3", "bytedance/sdxl-lightning-4step", "playgroundai/playground-v2.5-1024px-aesthetic"}
-    text_models = {"meta/meta-llama-3-70b-instruct", "mistralai/mixtral-8x7b-instruct-v0.1", "google-deepmind/gemma-2b-it"}
+    image_models = [
+        'black-forest-labs/flux-schnell',
+        'stability-ai/stable-diffusion-3',
+        'bytedance/sdxl-lightning-4step',
+        'playgroundai/playground-v2.5-1024px-aesthetic',
+    ]
 
+    models = text_models + image_models
+    
     model_aliases = {
+        "flux-schnell": "black-forest-labs/flux-schnell",
         "sd-3": "stability-ai/stable-diffusion-3",
         "sdxl": "bytedance/sdxl-lightning-4step",
         "playground-v2.5": "playgroundai/playground-v2.5-1024px-aesthetic",
         "llama-3-70b": "meta/meta-llama-3-70b-instruct",
         "mixtral-8x7b": "mistralai/mixtral-8x7b-instruct-v0.1",
         "gemma-2b": "google-deepmind/gemma-2b-it",
+        "llava-13b": "yorickvp/llava-13b",
+    }
+
+    model_versions = {
+        "meta/meta-llama-3-70b-instruct": "fbfb20b472b2f3bdd101412a9f70a0ed4fc0ced78a77ff00970ee7a2383c575d",
+        "mistralai/mixtral-8x7b-instruct-v0.1": "5d78bcd7a992c4b793465bcdcf551dc2ab9668d12bb7aa714557a21c1e77041c",
+        "google-deepmind/gemma-2b-it": "dff94eaf770e1fc211e425a50b51baa8e4cac6c39ef074681f9e39d778773626",
+        "yorickvp/llava-13b": "80537f9eead1a5bfa72d5ac6ea6414379be41d4d4f6679fd776e9535d1eb58bb",
+        'black-forest-labs/flux-schnell': "f2ab8a5bfe79f02f0789a146cf5e73d2a4ff2684a98c2b303d1e1ff3814271db",
+        'stability-ai/stable-diffusion-3': "527d2a6296facb8e47ba1eaf17f142c240c19a30894f437feee9b91cc29d8e4f",
+        'bytedance/sdxl-lightning-4step': "5f24084160c9089501c1b3545d9be3c27883ae2239b6f412990e82d4a6210f8f",
+        'playgroundai/playground-v2.5-1024px-aesthetic': "a45f82a1382bed5c7aeb861dac7c7d191b0fdf74d8d57c4a0e6ed7d4d0bf7d24",
     }
 
     @classmethod
@@ -77,84 +71,73 @@ class ReplicateHome(AsyncGeneratorProvider, ProviderModelMixin):
         cls,
         model: str,
         messages: Messages,
-        **kwargs: Any
-    ) -> Generator[Union[str, ImageResponse], None, None]:
-        yield await cls.create_async(messages[-1]["content"], model, **kwargs)
-
-    @classmethod
-    async def create_async(
-        cls,
-        prompt: str,
-        model: str,
-        api_key: Optional[str] = None,
-        proxy: Optional[str] = None,
-        timeout: int = 180,
-        version: Optional[str] = None,
-        extra_data: Dict[str, Any] = {},
-        **kwargs: Any
-    ) -> Union[str, ImageResponse]:
-        model = cls.get_model(model)  # Use the get_model method to resolve model name
+        proxy: str = None,
+        **kwargs
+    ) -> AsyncResult:
+        model = cls.get_model(model)
+        
         headers = {
-            'Accept-Encoding': 'gzip, deflate, br',
-            'Accept-Language': 'en-US',
-            'Connection': 'keep-alive',
-            'Origin': cls.url,
-            'Referer': f'{cls.url}/',
-            'Sec-Fetch-Dest': 'empty',
-            'Sec-Fetch-Mode': 'cors',
-            'Sec-Fetch-Site': 'same-site',
-            'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/119.0.0.0 Safari/537.36',
-            'sec-ch-ua': '"Google Chrome";v="119", "Chromium";v="119", "Not?A_Brand";v="24"',
-            'sec-ch-ua-mobile': '?0',
-            'sec-ch-ua-platform': '"macOS"',
+            "accept": "*/*",
+            "accept-language": "en-US,en;q=0.9",
+            "cache-control": "no-cache",
+            "content-type": "application/json",
+            "origin": "https://replicate.com",
+            "pragma": "no-cache",
+            "priority": "u=1, i",
+            "referer": "https://replicate.com/",
+            "sec-ch-ua": '"Not;A=Brand";v="24", "Chromium";v="128"',
+            "sec-ch-ua-mobile": "?0",
+            "sec-ch-ua-platform": '"Linux"',
+            "sec-fetch-dest": "empty",
+            "sec-fetch-mode": "cors",
+            "sec-fetch-site": "same-site",
+            "user-agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/128.0.0.0 Safari/537.36"
         }
-
-        if version is None:
-            version = random.choice(cls.versions.get(model, []))
-        if api_key is not None:
-            headers["Authorization"] = f"Bearer {api_key}"
-
-        async with StreamSession(
-            proxies={"all": proxy},
-            headers=headers,
-            timeout=timeout
-        ) as session:
+        
+        async with ClientSession(headers=headers) as session:
+            if model in cls.image_models:
+                prompt = messages[-1]['content'] if messages else ""
+            else:
+                prompt = format_prompt(messages)
+            
             data = {
-                "input": {
-                    "prompt": prompt,
-                    **extra_data
-                },
-                "version": version
+                "model": model,
+                "version": cls.model_versions[model],
+                "input": {"prompt": prompt},
             }
-            if api_key is None:
-                data["model"] = model
-                url = "https://homepage.replicate.com/api/prediction"
-            else:
-                url = "https://api.replicate.com/v1/predictions"
-            async with session.post(url, json=data) as response:
-                await raise_for_status(response)
+            
+            async with session.post(cls.api_endpoint, json=data, proxy=proxy) as response:
+                response.raise_for_status()
                 result = await response.json()
-            if "id" not in result:
-                raise ResponseError(f"Invalid response: {result}")
+                prediction_id = result['id']
+            
+            poll_url = f"https://homepage.replicate.com/api/poll?id={prediction_id}"
+            max_attempts = 30
+            delay = 5
+            for _ in range(max_attempts):
+                async with session.get(poll_url, proxy=proxy) as response:
+                    response.raise_for_status()
+                    try:
+                        result = await response.json()
+                    except ContentTypeError:
+                        text = await response.text()
+                        try:
+                            result = json.loads(text)
+                        except json.JSONDecodeError:
+                            raise ValueError(f"Unexpected response format: {text}")
 
-            while True:
-                if api_key is None:
-                    url = f"https://homepage.replicate.com/api/poll?id={result['id']}"
-                else:
-                    url = f"https://api.replicate.com/v1/predictions/{result['id']}"
-                async with session.get(url) as response:
-                    await raise_for_status(response)
-                    result = await response.json()
-                    if "status" not in result:
-                        raise ResponseError(f"Invalid response: {result}")
-                    if result["status"] == "succeeded":
-                        output = result['output']
-                        if model in cls.text_models:
-                            return ''.join(output) if isinstance(output, list) else output
-                        elif model in cls.image_models:
-                            images: List[Any] = output
-                            images = images[0] if len(images) == 1 else images
-                            return ImageResponse(images, prompt)
-                    elif result["status"] == "failed":
-                        raise ResponseError(f"Prediction failed: {result}")
-                    await asyncio.sleep(0.5)
+                    if result['status'] == 'succeeded':
+                        if model in cls.image_models:
+                            image_url = result['output'][0]
+                            yield ImageResponse(image_url, "Generated image")
+                            return
+                        else:
+                            for chunk in result['output']:
+                                yield chunk
+                        break
+                    elif result['status'] == 'failed':
+                        raise Exception(f"Prediction failed: {result.get('error')}")
+                await asyncio.sleep(delay)
+            
+            if result['status'] != 'succeeded':
+                raise Exception("Prediction timed out")
diff --git a/g4f/Provider/Rocks.py b/g4f/Provider/Rocks.py
deleted file mode 100644
index f44e0060..00000000
--- a/g4f/Provider/Rocks.py
+++ /dev/null
@@ -1,70 +0,0 @@
-import asyncio
-import json
-from aiohttp import ClientSession
-from ..typing import Messages, AsyncResult
-from .base_provider import AsyncGeneratorProvider
-
-class Rocks(AsyncGeneratorProvider):
-    url = "https://api.airforce"
-    api_endpoint = "/chat/completions"
-    supports_message_history = True
-    supports_gpt_35_turbo = True
-    supports_gpt_4 = True
-    supports_stream = True
-    supports_system_message = True
-    working = True
-
-    @classmethod
-    async def create_async_generator(
-        cls,
-        model: str,
-        messages: Messages,
-        proxy: str = None,
-        **kwargs
-    ) -> AsyncResult:
-        payload = {"messages":messages,"model":model,"max_tokens":4096,"temperature":1,"top_p":1,"stream":True}
-
-        headers = {
-            "Accept": "application/json",
-            "Accept-Encoding": "gzip, deflate, br, zstd",
-            "Accept-Language": "en-US,en;q=0.9",
-            "Authorization": "Bearer missing api key",
-            "Origin": "https://llmplayground.net",
-            "Referer": "https://llmplayground.net/",
-            "Sec-Fetch-Dest": "empty",
-            "Sec-Fetch-Mode": "cors",
-            "Sec-Fetch-Site": "same-origin",
-            "User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/127.0.0.0 Safari/537.36",
-        }
-
-        async with ClientSession() as session:
-            async with session.post(
-                f"{cls.url}{cls.api_endpoint}",
-                json=payload,
-                proxy=proxy,
-                headers=headers
-            ) as response:
-                response.raise_for_status()
-                last_chunk_time = asyncio.get_event_loop().time()
-                
-                async for line in response.content:
-                    current_time = asyncio.get_event_loop().time()
-                    if current_time - last_chunk_time > 5:
-                        return
-                    
-                    if line.startswith(b"\n"):
-                        pass
-                    elif "discord.com/invite/" in line.decode() or "discord.gg/" in line.decode():
-                        pass # trolled
-                    elif line.startswith(b"data: "):
-                        try:
-                            line = json.loads(line[6:])
-                        except json.JSONDecodeError:
-                            continue
-                        chunk = line["choices"][0]["delta"].get("content")
-                        if chunk:
-                            yield chunk
-                            last_chunk_time = current_time
-                    else:
-                        raise Exception(f"Unexpected line: {line}")
-                return
\ No newline at end of file
diff --git a/g4f/Provider/Snova.py b/g4f/Provider/Snova.py
index 76dfac40..53d8f0bd 100644
--- a/g4f/Provider/Snova.py
+++ b/g4f/Provider/Snova.py
@@ -24,10 +24,9 @@ class Snova(AsyncGeneratorProvider, ProviderModelMixin):
         'Meta-Llama-3.1-70B-Instruct',
         'Meta-Llama-3.1-405B-Instruct',
         'Samba-CoE',
-        'ignos/Mistral-T5-7B-v1',
+        'ignos/Mistral-T5-7B-v1', # Error with the answer
         'v1olet/v1olet_merged_dpo_7B',
         'macadeliccc/WestLake-7B-v2-laser-truthy-dpo',
-        'cookinai/DonutLM-v1',
     ]
     
     model_aliases = {
@@ -40,7 +39,6 @@ class Snova(AsyncGeneratorProvider, ProviderModelMixin):
         "samba-coe-v0.1": "Samba-CoE",
         "v1olet-merged-7b": "v1olet/v1olet_merged_dpo_7B",
         "westlake-7b-v2": "macadeliccc/WestLake-7B-v2-laser-truthy-dpo",
-        "donutlm-v1": "cookinai/DonutLM-v1",
     }
 
     @classmethod
diff --git a/g4f/Provider/__init__.py b/g4f/Provider/__init__.py
index a9a815ea..e4c85d6a 100644
--- a/g4f/Provider/__init__.py
+++ b/g4f/Provider/__init__.py
@@ -13,6 +13,7 @@ from .AI365VIP         import AI365VIP
 from .Allyfy           import Allyfy
 from .AiChatOnline     import AiChatOnline
 from .AiChats          import AiChats
+from .Airforce         import Airforce
 from .Aura             import Aura
 from .Bing             import Bing
 from .BingCreateImages import BingCreateImages
@@ -28,7 +29,6 @@ from .DDG              import DDG
 from .DeepInfra        import DeepInfra
 from .DeepInfraImage   import DeepInfraImage
 from .FlowGpt          import FlowGpt
-from .FluxAirforce     import FluxAirforce
 from .Free2GPT         import Free2GPT
 from .FreeChatgpt      import FreeChatgpt
 from .FreeGpt          import FreeGpt
@@ -41,7 +41,6 @@ from .HuggingFace      import HuggingFace
 from .Koala            import Koala
 from .Liaobots         import Liaobots
 from .LiteIcoding      import LiteIcoding
-from .Llama            import Llama
 from .Local            import Local
 from .MagickPen        import MagickPen
 from .MetaAI           import MetaAI
@@ -51,11 +50,11 @@ from .Ollama           import Ollama
 from .PerplexityLabs   import PerplexityLabs
 from .Pi               import Pi
 from .Pizzagpt         import Pizzagpt
+from .Prodia           import Prodia
 from .Reka             import Reka
 from .Snova            import Snova
 from .Replicate        import Replicate
 from .ReplicateHome    import ReplicateHome
-from .Rocks            import Rocks
 from .TeachAnything    import TeachAnything
 from .TwitterBio       import TwitterBio
 from .Upstage          import Upstage
diff --git a/g4f/Provider/bing/conversation.py b/g4f/Provider/bing/conversation.py
index a4195fa4..b5c237f9 100644
--- a/g4f/Provider/bing/conversation.py
+++ b/g4f/Provider/bing/conversation.py
@@ -33,9 +33,9 @@ async def create_conversation(session: StreamSession, headers: dict, tone: str)
     Conversation: An instance representing the created conversation.
     """
     if tone == "Copilot":
-        url = "https://copilot.microsoft.com/turing/conversation/create?bundleVersion=1.1690.0"
+        url = "https://copilot.microsoft.com/turing/conversation/create?bundleVersion=1.1809.0"
     else:
-        url = "https://www.bing.com/turing/conversation/create?bundleVersion=1.1690.0"
+        url = "https://www.bing.com/turing/conversation/create?bundleVersion=1.1809.0"
     async with session.get(url, headers=headers) as response:
         if response.status == 404:
             raise RateLimitError("Response 404: Do less requests and reuse conversations")
@@ -90,4 +90,4 @@ async def delete_conversation(session: StreamSession, conversation: Conversation
             response = await response.json()
             return response["result"]["value"] == "Success"
     except:
-        return False
\ No newline at end of file
+        return False
diff --git a/g4f/Provider/selenium/AItianhuSpace.py b/g4f/Provider/selenium/AItianhuSpace.py
deleted file mode 100644
index 4c438e3b..00000000
--- a/g4f/Provider/selenium/AItianhuSpace.py
+++ /dev/null
@@ -1,116 +0,0 @@
-from __future__ import annotations
-
-import time
-import random
-
-from ...typing import CreateResult, Messages
-from ..base_provider import AbstractProvider
-from ..helper import format_prompt, get_random_string
-from ...webdriver import WebDriver, WebDriverSession, element_send_text
-from ... import debug
-
-class AItianhuSpace(AbstractProvider):
-    url = "https://chat3.aiyunos.top/"
-    working = True
-    supports_stream = True
-    supports_gpt_35_turbo = True
-    _domains = ["aitianhu.com", "aitianhu1.top"]
-
-    @classmethod
-    def create_completion(
-        cls,
-        model: str,
-        messages: Messages,
-        stream: bool,
-        domain: str = None,
-        proxy: str = None,
-        timeout: int = 120,
-        webdriver: WebDriver = None,
-        headless: bool = True,
-        **kwargs
-    ) -> CreateResult:
-        if not model:
-            model = "gpt-3.5-turbo"
-        if not domain:
-            rand = get_random_string(6)
-            domain = random.choice(cls._domains)
-            domain = f"{rand}.{domain}"
-        if debug.logging:
-            print(f"AItianhuSpace | using domain: {domain}")
-        url = f"https://{domain}"
-        prompt = format_prompt(messages)
-
-        with WebDriverSession(webdriver, "", headless=headless, proxy=proxy) as driver:
-            from selenium.webdriver.common.by import By
-            from selenium.webdriver.support.ui import WebDriverWait
-            from selenium.webdriver.support import expected_conditions as EC
-
-            wait = WebDriverWait(driver, timeout)
-
-            # Bypass devtools detection
-            driver.get("https://blank.page/")
-            wait.until(EC.visibility_of_element_located((By.ID, "sheet")))
-            driver.execute_script(f"""
-    document.getElementById('sheet').addEventListener('click', () => {{
-        window.open(arguments[0]);
-    }});
-            """, url)
-            driver.find_element(By.ID, "sheet").click()
-            time.sleep(10)
-
-            original_window = driver.current_window_handle
-            for window_handle in driver.window_handles:
-                if window_handle != original_window:
-                    driver.close()
-                    driver.switch_to.window(window_handle)
-                    break
-
-            # Wait for page load
-            wait.until(EC.visibility_of_element_located((By.CSS_SELECTOR, "textarea.n-input__textarea-el")))
-
-            # Register hook in XMLHttpRequest
-            script = """
-const _http_request_open = XMLHttpRequest.prototype.open;
-window._last_message = window._message = "";
-window._loadend = false;
-XMLHttpRequest.prototype.open = function(method, url) {
-    if (url == "/api/chat-process") {
-        this.addEventListener("progress", (event) => {
-            const lines = this.responseText.split("\\n");
-            try {
-                window._message = JSON.parse(lines[lines.length-1])["text"];
-            } catch(e) { }
-        });
-        this.addEventListener("loadend", (event) => {
-            window._loadend = true;
-        });
-    }
-    return _http_request_open.call(this, method, url);
-}
-"""
-            driver.execute_script(script)
-
-            # Submit prompt
-            element_send_text(driver.find_element(By.CSS_SELECTOR, "textarea.n-input__textarea-el"), prompt)
-
-            # Read response
-            while True:
-                chunk = driver.execute_script("""
-if (window._message && window._message != window._last_message) {
-    try {
-        return window._message.substring(window._last_message.length);
-    } finally {
-        window._last_message = window._message;
-    }
-}
-if (window._loadend) {
-    return null;
-}
-return "";
-""")
-                if chunk:
-                    yield chunk
-                elif chunk != "":
-                    break
-                else:
-                    time.sleep(0.1)
\ No newline at end of file
diff --git a/g4f/Provider/selenium/__init__.py b/g4f/Provider/selenium/__init__.py
index 9a020460..1b801725 100644
--- a/g4f/Provider/selenium/__init__.py
+++ b/g4f/Provider/selenium/__init__.py
@@ -1,4 +1,3 @@
-from .AItianhuSpace import AItianhuSpace
 from .MyShell import MyShell
 from .PerplexityAi import PerplexityAi
 from .Phind import Phind