[skip ci] Update nightly support matrices

Signed-off-by: Buildkite Bot <buildkite-bot@users.noreply.github.com>

Author: buildkite-bot

support_matrices/nightly/feature_support_matrix.csv

@@ -1,12 +1,12 @@
 Feature,CorrectnessTest,PerformanceTest
 "Chunked Prefill",✅,✅
 "Collective Communication Matmul",✅,N/A
-"JAX-Path Qxix Quantization",N/A,❌
-"Multimodal Inputs",✅,❌
+"JAX-Path Qxix Quantization",N/A,✅
+"Multimodal Inputs",✅,✅
 "Prefix Caching",✅,✅
-"Quantized Matmul Attention and KV Cache",N/A,❌
+"Quantized Matmul Attention and KV Cache",N/A,✅
 "Ragged Paged Attention V3",✅,✅
 "Single Program Multi Data",✅,✅
-"Speculative Decoding: Ngram",❌,N/A
-"Structured Decoding",❌,N/A
+"Speculative Decoding: Ngram",✅,✅
+"Structured Decoding",✅,N/A
 "async scheduler",❌,N/A

support_matrices/nightly/model_support_matrix.csv

@@ -1,8 +1,8 @@
 Model,UnitTest,IntegrationTest,Benchmark
 "meta-llama/Llama-3.3-70B-Instruct",❌,N/A,N/A
-"Qwen/Qwen3-4B",✅,❌,N/A
+"Qwen/Qwen3-4B",✅,✅,✅
 "Qwen/Qwen2.5-VL-7B-Instruct",✅,❌,N/A
 "google/gemma-3-27b-it",❌,N/A,N/A
 "Qwen/Qwen3-32B",✅,❌,N/A
 "meta-llama/Llama-3.1-8B-Instruct",❌,N/A,N/A
-"Qwen/Qwen3-30B-A3B",✅,❌,N/A
+"Qwen/Qwen3-30B-A3B",✅,✅,❌