add Groq LLaMA3 70B support llama3-70b-8192 (#1685)

mckaywrigley · Apr 22, 2024 · d5a2d28 · d5a2d28 · chikosan · Apr 23, 2024
1 parent 38eeadf
commit d5a2d28
Show file tree

Hide file tree

Showing 3 changed files with 23 additions and 1 deletion.
diff --git a/lib/chat-setting-limits.ts b/lib/chat-setting-limits.ts
@@ -87,6 +87,12 @@ export const CHAT_SETTING_LIMITS: Record<LLMID, ChatSettingLimits> = {
     MAX_TOKEN_OUTPUT_LENGTH: 4096,
     MAX_CONTEXT_LENGTH: 4096
   },
+  "llama3-70b-8192": {
+    MIN_TEMPERATURE: 0.0,
+    MAX_TEMPERATURE: 1.0,
+    MAX_TOKEN_OUTPUT_LENGTH: 8192,
+    MAX_CONTEXT_LENGTH: 8192
+  },
   "mixtral-8x7b-32768": {
     MIN_TEMPERATURE: 0.0,
     MAX_TEMPERATURE: 1.0,

diff --git a/lib/models/llm/groq-llm-list.ts b/lib/models/llm/groq-llm-list.ts
@@ -17,6 +17,21 @@ const LLaMA2_70B: LLM = {
   }
 }
 
+const LLaMA3_70B: LLM = {
+  modelId: "llama3-70b-8192",
+  modelName: "LLaMA3-70b-chat",
+  provider: "groq",
+  hostedId: "llama3-70b-4096",
+  platformLink: GROQ_PLATORM_LINK,
+  imageInput: false,
+  pricing: {
+    currency: "USD",
+    unit: "1M tokens",
+    inputCost: 0.78,
+    outputCost: 0.9
+  }
+}
+
 const MIXTRAL_8X7B: LLM = {
   modelId: "mixtral-8x7b-32768",
   modelName: "Mixtral-8x7b-Instruct-v0.1",
@@ -32,4 +47,4 @@ const MIXTRAL_8X7B: LLM = {
   }
 }
 
-export const GROQ_LLM_LIST: LLM[] = [LLaMA2_70B, MIXTRAL_8X7B]
+export const GROQ_LLM_LIST: LLM[] = [LLaMA2_70B, LLaMA3_70B, MIXTRAL_8X7B]
diff --git a/types/llms.ts b/types/llms.ts
@@ -37,6 +37,7 @@ export type MistralLLMID =
 
 export type GroqLLMID =
   | "llama2-70b-4096" // LLaMA2-70b
+  | "llama3-70b-8192" // LLaMA3-70b
   | "mixtral-8x7b-32768" // Mixtral-8x7b
 
 // Perplexity Models (UPDATED 1/31/24)