perf(agentx): refresh MI355X MiniMax M3

中文：使用新版 AgentX 测试框架刷新 MI355X MiniMax M3 基准测试入口。

Author: cquil11

perf-changelog.yaml

@@ -5355,3 +5355,9 @@
     - "Apply the accuracy-gated Kimi-K2.5 MXFP4 settings: tuned AITER MXFP4 MoE, fused shared experts, FP8 KV cache, block size 16, 16384 batched tokens, 512 sequences, async scheduling, gpu-memory-utilization 0.85 (headroom for CUDA-graph capture on MI355X), and the AITER BF16 GEMM path"
     - "Extend the TP4 and TP8 8k1k concurrency sweep from 64 to 128 (1k1k deprecated per #2263)"
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2213
+
+- config-keys:
+    - minimaxm3-fp4-mi355x-vllm-agentic
+  description:
+    - "Initial MI355X MiniMax M3 with new AgentX harness"
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2458