llm-simple-0.1.0.1: model-catalog.json
[
{
"modelConfigName": "gpt_4_1",
"providerName": "openai",
"modelName": "gpt-4.1-2025-04-14",
"pricing": {
"pricePerMillionInput": 2.0,
"pricePerMillionOutput": 8.0
},
"maxTokens": 1024,
"temperature": 0.5,
"requestTimeout": 10000,
"throttleDelay": 1000,
"retryCount": 3,
"jitterBackoff": 1000
},
{
"modelConfigName": "llama_3_2",
"providerName": "ollama",
"modelName": "llama3.2:latest",
"pricing": {
"pricePerMillionInput": 0.0,
"pricePerMillionOutput": 0.0
},
"maxTokens": 1024,
"temperature": 0.5,
"requestTimeout": 3000,
"throttleDelay": 1000,
"retryCount": 3,
"jitterBackoff": 1000
},
{
"modelConfigName": "mistral",
"providerName": "ollama",
"modelName": "mistral:latest",
"pricing": {
"pricePerMillionInput": 0.0,
"pricePerMillionOutput": 0.0
},
"maxTokens": 1024,
"temperature": 0.5,
"requestTimeout": 30000,
"throttleDelay": 1000,
"retryCount": 3,
"jitterBackoff": 1000
},
{
"modelConfigName": "gemini_2_5_flash",
"providerName": "gemini",
"modelName": "gemini-2.5-flash",
"pricing": {
"pricePerMillionInput": 0.1,
"pricePerMillionOutput": 0.4
},
"maxTokens": 1024,
"temperature": 0.5,
"requestTimeout": 10000,
"throttleDelay": 1000,
"retryCount": 3,
"jitterBackoff": 1000
},
{
"modelConfigName": "gemini_lite",
"providerName": "gemini",
"modelName": "gemini-3.1-flash-lite",
"pricing": {
"pricePerMillionInput": 0.1,
"pricePerMillionOutput": 0.4
},
"maxTokens": 1024,
"temperature": 0.5,
"requestTimeout": 10000,
"throttleDelay": 1000,
"retryCount": 3,
"jitterBackoff": 1000
},
{
"modelConfigName": "haiku_4_5",
"providerName": "claude",
"modelName": "claude-haiku-4-5-20251001",
"pricing": {
"pricePerMillionInput": 1,
"pricePerMillionOutput": 5
},
"maxTokens": 1024,
"temperature": 0.5,
"requestTimeout": 30000,
"throttleDelay": 5000,
"retryCount": 3,
"jitterBackoff": 5000
},
{
"modelConfigName": "deepseek4flash",
"providerName": "deepseek",
"modelName": "deepseek-v4-flash",
"pricing": {
"pricePerMillionInput": 0.14,
"pricePerMillionOutput": 0.28
},
"maxTokens": 1024,
"temperature": 0.5,
"requestTimeout": 20000,
"throttleDelay": 3000,
"retryCount": 3,
"jitterBackoff": 1000
}
]