add qwen3.5 and fix link to llama

Author: robmsmt

frontend/src/lib/modelMetrics.ts

@@ -8,9 +8,10 @@ const modelEngines: Record<string, Engine> = {
   "zai-org/GLM-4.7-Flash": "sglang",
   "Snowflake/snowflake-arctic-embed-l-v2.0": "vllm",
   "cais/HarmBench-Llama-2-13b-cls": "vllm",
-  "meta-llama/Llama-3.3-70B-Instruct": "vllm",
+  "meta-llama/Llama-3.3-70B-Instruct": "sglang",
   "meta-llama/Llama-Guard-4-12B": "vllm",
   "swiss-ai/Apertus-70B-Instruct-2509": "vllm",
+  "Qwen/Qwen3.5-27B": "sglang",
 };
 
 /**