salman/igny8
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add support for GPT-5.1 and GPT-5.2: update token limits and pricing

Browse Source
This commit is contained in:
IGNY8 VPS (Salman)
2025-12-17 11:11:11 +00:00
parent e43f8553b6
commit 62e55389f9
7 changed files with 40 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
backend/igny8_core/ai/ai_core.py
View File

@@ -239,8 +239,12 @@ class AICore:
'temperature': temperature,
}
# GPT-5.1 and GPT-5.2 use max_completion_tokens instead of max_tokens
if max_tokens:
body_data['max_tokens'] = max_tokens
if active_model in ['gpt-5.1', 'gpt-5.2']:
body_data['max_completion_tokens'] = max_tokens
else:
body_data['max_tokens'] = max_tokens
if response_format:
body_data['response_format'] = response_format

4
backend/igny8_core/ai/constants.py
View File

@@ -6,6 +6,8 @@ MODEL_RATES = {
'gpt-4.1': {'input': 2.00, 'output': 8.00},
'gpt-4o-mini': {'input': 0.15, 'output': 0.60},
'gpt-4o': {'input': 2.50, 'output': 10.00},
'gpt-5.1': {'input': 1.25, 'output': 10.00},
'gpt-5.2': {'input': 1.75, 'output': 14.00},
}
# Image model pricing (per image) - EXACT from reference plugin
@@ -33,7 +35,7 @@ VALID_SIZES_BY_MODEL = {
DEFAULT_AI_MODEL = 'gpt-4.1'
# JSON mode supported models
JSON_MODE_MODELS = ['gpt-4o', 'gpt-4o-mini', 'gpt-4-turbo-preview']
JSON_MODE_MODELS = ['gpt-4o', 'gpt-4o-mini', 'gpt-4-turbo-preview', 'gpt-5.1', 'gpt-5.2']
# Debug mode - controls console logging
# Set to False in production to disable verbose logging

6
backend/igny8_core/ai/settings.py
View File

@@ -99,8 +99,10 @@ def get_model_config(function_name: str, account) -> Dict[str, Any]:
# MODEL_RATES not available - skip validation
pass
# Get max_tokens and temperature from config (standardized to 8192)
max_tokens = config.get('max_tokens', 8192) # Standardized across entire codebase
# Get max_tokens and temperature from config (standardized to 8192, 16384 for GPT-5.x)
# GPT-5.1 and GPT-5.2 use 16384 max_tokens by default
default_max_tokens = 16384 if model in ['gpt-5.1', 'gpt-5.2'] else 8192
max_tokens = config.get('max_tokens', default_max_tokens)
temperature = config.get('temperature', 0.7) # Reasonable default
# Build response format based on model (JSON mode for supported models)

2
backend/igny8_core/api/tests/test_ai_framework.py
View File

@@ -140,7 +140,7 @@ class GetModelConfigTestCase(TestCase):
def test_get_model_config_json_mode_models(self):
"""Test get_model_config() sets response_format for JSON mode models"""
json_models = ['gpt-4o', 'gpt-4o-mini', 'gpt-4-turbo-preview']
json_models = ['gpt-4o', 'gpt-4o-mini', 'gpt-4-turbo-preview', 'gpt-5.1', 'gpt-5.2']
for model in json_models:
IntegrationSettings.objects.filter(account=self.account).delete()

13
backend/igny8_core/modules/system/integration_views.py
View File

@@ -191,8 +191,10 @@ class IntegrationSettingsViewSet(viewsets.ViewSet):
# Get model from config or use default (reference plugin: get_option('igny8_model', 'gpt-4.1'))
model = (config or {}).get('model', 'gpt-4.1') if config else 'gpt-4.1'
# Get max_tokens from config or use default
max_tokens = (config or {}).get('max_tokens', 8192) if config else 8192
# Get max_tokens from config or use default based on model
# GPT-5.1 and GPT-5.2 use 16384 max_tokens by default
default_max_tokens = 16384 if model in ['gpt-5.1', 'gpt-5.2'] else 8192
max_tokens = (config or {}).get('max_tokens', default_max_tokens) if config else default_max_tokens
# Check if test with response is requested (reference plugin: $with_response parameter)
with_response = (config or {}).get('with_response', False) if config else False
@@ -207,10 +209,15 @@ class IntegrationSettingsViewSet(viewsets.ViewSet):
'content': 'Reply with exactly: OK! Ping Received'
}
],
'max_tokens': max_tokens,
'temperature': 0.7,
}
# GPT-5.1 and GPT-5.2 use max_completion_tokens instead of max_tokens
if model in ['gpt-5.1', 'gpt-5.2']:
request_body['max_completion_tokens'] = max_tokens
else:
request_body['max_tokens'] = max_tokens
try:
response = requests.post(
'https://api.openai.com/v1/chat/completions',

22
backend/igny8_core/utils/ai_processor.py
View File

@@ -174,8 +174,12 @@ class AIProcessor:
}
# Add max_tokens if specified (reference plugin uses 4000 default)
# GPT-5.1 and GPT-5.2 use max_completion_tokens instead of max_tokens
if max_tokens:
body_data['max_tokens'] = max_tokens
if model in ['gpt-5.1', 'gpt-5.2']:
body_data['max_completion_tokens'] = max_tokens
else:
body_data['max_tokens'] = max_tokens
# Add response_format if specified (for JSON mode)
if response_format:
@@ -1131,15 +1135,17 @@ Make sure each prompt is detailed enough for image generation, describing the vi
# Call OpenAI with JSON response format - Steps 8-10 happen in _call_openai
# Use the active model from integration settings (self.default_model)
# Check if model supports JSON mode
json_models = ['gpt-4o', 'gpt-4o-mini', 'gpt-4-turbo-preview']
active_model = self.default_model # This is the active model from IntegrationSettings
response_format = {'type': 'json_object'} if active_model in json_models else None
response_format = {'type': 'json_object'} if active_model in JSON_MODE_MODELS else None
# GPT-5.1 and GPT-5.2 use 16384 max_tokens
max_tokens = 16384 if active_model in ['gpt-5.1', 'gpt-5.2'] else 8192
try:
result = self._call_openai(
prompt,
model=active_model, # Explicitly pass to ensure consistency
max_tokens=8192,
max_tokens=max_tokens,
temperature=0.7,
response_format=response_format,
response_steps=response_steps
@@ -1339,14 +1345,16 @@ Make sure each prompt is detailed enough for image generation, describing the vi
# Use the active model from integration settings (self.default_model)
# Check if model supports JSON mode
json_models = ['gpt-4o', 'gpt-4o-mini', 'gpt-4-turbo-preview']
active_model = self.default_model # This is the active model from IntegrationSettings
response_format = {'type': 'json_object'} if active_model in json_models else None
response_format = {'type': 'json_object'} if active_model in JSON_MODE_MODELS else None
# GPT-5.1 and GPT-5.2 use 16384 max_tokens
max_tokens = 16384 if active_model in ['gpt-5.1', 'gpt-5.2'] else 8192
result = self._call_openai(
prompt,
model=active_model, # Explicitly pass to ensure consistency
max_tokens=8192,
max_tokens=max_tokens,
temperature=0.7,
response_format=response_format
)