Add test case for Qwen3N (#2532)

<!-- .github/pull_request_template.md -->

## 📌 Description
Add test case for Qwen3N, and Qwen3.5 according to
vllm-project/vllm#34131
<!-- What does this PR do? Briefly describe the changes and why they’re
needed. -->

## 🔍 Related Issues

<!-- Link any related issues here -->

## 🚀 Pull Request Checklist

Thank you for contributing to FlashInfer! Before we review your pull
request, please make sure the following items are complete.

### ✅ Pre-commit Checks

- [x] I have installed `pre-commit` by running `pip install pre-commit`
(or used your preferred method).
- [x] I have installed the hooks with `pre-commit install`.
- [x] I have run the hooks manually with `pre-commit run --all-files`
and fixed any reported issues.

> If you are unsure about how to set up `pre-commit`, see [the
pre-commit documentation](https://pre-commit.com/).

## 🧪 Tests

- [x] Tests have been added or updated as needed.
- [ ] All tests are passing (`unittest`, etc.).

## Reviewer Notes

<!-- Optional: anything you'd like reviewers to focus on, concerns, etc.
-->


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **Tests**
* Expanded test coverage by adding additional head-configuration cases
across multiple test scenarios to improve reliability and catch more
edge cases.
* No changes to test logic or public interfaces; only parameterized
inputs were extended.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Author: samuellees

tests/gdn/test_prefill_delta_rule.py

@@ -144,7 +144,16 @@ def _test_prefill_kernel(
 @pytest.mark.parametrize("head_size", [128])
 @pytest.mark.parametrize(
     "num_q_heads, num_k_heads, num_v_heads",
-    [(1, 1, 1), (4, 1, 1), (3, 3, 3), (6, 2, 2), (1, 1, 2), (2, 2, 4)],
+    [
+        (1, 1, 1),
+        (4, 1, 1),
+        (3, 3, 3),
+        (6, 2, 2),
+        (1, 1, 2),
+        (2, 2, 4),
+        (16, 16, 32),
+        (16, 16, 64),
+    ],
 )
 @pytest.mark.parametrize("seq_lens", [[64], [128], [256], [256, 256], [64, 128, 512]])
 @pytest.mark.parametrize("block_size", [64])
@@ -186,7 +195,16 @@ def test_prefill_kernel_basic(
 @pytest.mark.parametrize("head_size", [128])
 @pytest.mark.parametrize(
     "num_q_heads, num_k_heads, num_v_heads",
-    [(1, 1, 1), (4, 1, 1), (3, 3, 3), (6, 2, 2), (1, 1, 2), (2, 2, 4)],
+    [
+        (1, 1, 1),
+        (4, 1, 1),
+        (3, 3, 3),
+        (6, 2, 2),
+        (1, 1, 2),
+        (2, 2, 4),
+        (16, 16, 32),
+        (16, 16, 64),
+    ],
 )
 @pytest.mark.parametrize(
     "seq_lens",
@@ -390,7 +408,8 @@ def concat_varlen(t1, cu_seq_lens1, t2, cu_seq_lens2):
 @pytest.mark.parametrize("scale", [1.0, "auto"])
 @pytest.mark.parametrize("head_size", [128])
 @pytest.mark.parametrize(
-    "num_q_heads, num_k_heads, num_v_heads", [(6, 2, 2), (2, 2, 4)]
+    "num_q_heads, num_k_heads, num_v_heads",
+    [(6, 2, 2), (2, 2, 4), (16, 16, 32), (16, 16, 64)],
 )
 @pytest.mark.parametrize(
     "seq_lens1, seq_lens2",