OpenAi and Groq Api and LM Studio API paths are similar, to get llama 3.1 model in this extension,
make the following changes, and use groq key as openai apikey.
try this prompt
wrap your tools calls like this <tool_call>your content here</tool_call> avoid <tool_response> <function_call> <function> for tool calling
The tool calls are working fine, but the model is fragile compared to the anthropic/claude-3.5-sonnet , with an 8k maxtokens limit.
OpenAi and Groq Api and LM Studio API paths are similar, to get llama 3.1 model in this extension, make the following changes, and use groq key as openai apikey.
file : openRouter.ts: baseURL: "https://api.groq.com/openai/v1",
file : api.ts: export const openRouterDefaultModelId: OpenRouterModelId = "llama3-groq-70b-8192-tool-use-preview"
export const openRouterModels = { "llama3-groq-70b-8192-tool-use-preview": { maxTokens: 8192, supportsImages: false, supportsPromptCache:false, inputPrice: 0.0, outputPrice: 0.0, },"mixtral-8x7b-32768": { maxTokens: 8192, supportsImages: false, supportsPromptCache:false, inputPrice: 0.0, outputPrice: 0.0, },
in setting panel, custom instructions
try this prompt
wrap your tools calls like this <tool_call>your content here</tool_call> avoid <tool_response> <function_call> <function> for tool calling
The tool calls are working fine, but the model is fragile compared to the anthropic/claude-3.5-sonnet , with an 8k maxtokens limit.