@@ -20,32 +20,27 @@ export function groqLLMRegistry(): Record<string, () => LLM> {
2020// https://groq.com/pricing/
2121// https://console.groq.com/docs/models
2222
23- // Qwen3 32B 131khttps://console.groq.com/docs/model/qwen3-32b
24- // 16,384 max output tokens
23+ // https://console.groq.com/docs/model/qwen/qwen3-32b
2524export function groqQwen3_32b ( ) : LLM {
26- return new GroqLLM ( 'Qwen3 32b (Groq)' , 'qwen/qwen3-32b' , 131_072 , costPerMilTokens ( 0.29 , 0.59 ) ) ;
25+ return new GroqLLM ( 'Qwen3 32b (Groq)' , 'qwen/qwen3-32b' , 131_072 , 40_960 , costPerMilTokens ( 0.29 , 0.59 ) ) ;
2726}
2827
28+ // https://console.groq.com/docs/model/meta-llama/llama-4-scout-17b-16e-instruct
2929export function groqLlama4_Scout ( ) : LLM {
30- return new GroqLLM ( 'Llama4 Scout (Groq)' , 'meta-llama/llama-4-scout-17b-16e-instruct' , 131_072 , costPerMilTokens ( 0.11 , 0.34 ) ) ;
30+ return new GroqLLM ( 'Llama4 Scout (Groq)' , 'meta-llama/llama-4-scout-17b-16e-instruct' , 131_072 , 8_192 , costPerMilTokens ( 0.11 , 0.34 ) ) ;
3131}
3232
33+ // https://console.groq.com/docs/model/moonshotai/kimi-k2-instruct-0905
3334export function groqKimiK2 ( ) : LLM {
34- return new GroqLLM (
35- 'Kimi K2 (Groq)' ,
36- 'moonshotai/kimi-k2-instruct' ,
37- // 16,384 max output tokens (from official Groq documentation)
38- 16384 ,
39- costPerMilTokens ( 1.0 , 3.0 ) ,
40- ) ;
35+ return new GroqLLM ( 'Kimi K2 (Groq)' , 'moonshotai/kimi-k2-instruct' , 262_144 , 16_384 , costPerMilTokens ( 1.0 , 3.0 ) ) ;
4136}
4237
4338/**
4439 * https://wow.groq.com/
4540 */
4641export class GroqLLM extends AiLLM < GroqProvider > {
47- constructor ( displayName : string , model : string , maxOutputTokens : number , calculateCosts : LlmCostFunction ) {
48- super ( { displayName, service : GROQ_SERVICE , modelId : model , maxInputTokens : maxOutputTokens , calculateCosts } ) ;
42+ constructor ( displayName : string , model : string , maxInputTokens : number , maxOutputTokens : number , calculateCosts : LlmCostFunction ) {
43+ super ( { displayName, service : GROQ_SERVICE , modelId : model , maxInputTokens, maxOutputTokens, calculateCosts } ) ;
4944 }
5045
5146 override aiModel ( ) : LanguageModelV2 {
0 commit comments