diff --git a/python/tokenspeed/runtime/execution/drafter/eagle.py b/python/tokenspeed/runtime/execution/drafter/eagle.py
index 81208140d..d51618813 100644
--- a/python/tokenspeed/runtime/execution/drafter/eagle.py
+++ b/python/tokenspeed/runtime/execution/drafter/eagle.py
@@ -36,6 +36,7 @@
     CaptureHiddenMode,
     ForwardMode,
 )
+from tokenspeed.runtime.models.base import BaseCausalLM
 from tokenspeed.runtime.multimodal.inputs import maybe_substitute_mm_pad
 from tokenspeed.runtime.utils import get_colorful_logger
 from tokenspeed.runtime.utils.nvtx import nvtx_range
@@ -226,9 +227,16 @@ def _run_first_step(
         )
         input_ids = maybe_substitute_mm_pad(input_ids, self.mm_pad_substitute_id)
         draft_first_step_reduce = forward_mode.is_decode()
-        if draft_first_step_reduce and self.attn_backend.support_kv_cache_prewrite:
-            # Trim seq_lens by rejected-draft count so the sliced decode
-            # query does not attend to dead positions.
+
+        # TODO: remove the isinstance/flag gate together with pre_attention_trim
+        # once Qwen NextN and DeepSeek V3 NextN also pre-slice q.
+        draft_model = self.draft_model_runner.model
+        if (
+            draft_first_step_reduce
+            and self.attn_backend.support_kv_cache_prewrite
+            and isinstance(draft_model, BaseCausalLM)
+            and draft_model.pre_attention_trim
+        ):
             correction = (self.spec_num_tokens - draft_input.accept_lengths).to(
                 self.draft_seq_lens_buf.dtype
             )
diff --git a/python/tokenspeed/runtime/models/base/causal_lm.py b/python/tokenspeed/runtime/models/base/causal_lm.py
index 24b1d05fa..1cd84880c 100644
--- a/python/tokenspeed/runtime/models/base/causal_lm.py
+++ b/python/tokenspeed/runtime/models/base/causal_lm.py
@@ -44,6 +44,10 @@ class BaseCausalLM(nn.Module):
 
     model_cls: type[BaseTransformerModel]
 
+    # TODO: temporary; remove in the follow-up refactoring that extends
+    # pre-attn q-slice to Qwen NextN and DeepSeek V3 NextN.
+    pre_attention_trim: bool = False
+
     def __init__(
         self,
         config: PretrainedConfig,
diff --git a/python/tokenspeed/runtime/models/llama_eagle3.py b/python/tokenspeed/runtime/models/llama_eagle3.py
index a64745e47..9c160f53b 100644
--- a/python/tokenspeed/runtime/models/llama_eagle3.py
+++ b/python/tokenspeed/runtime/models/llama_eagle3.py
@@ -583,6 +583,11 @@ class LlamaForCausalLMEagle3(BaseCausalLM):
 
     model_cls = Eagle3LlamaModel
 
+    # Eagle3 catch-up pre-slices q to active row before attn; trim must pair.
+    # TODO: remove together with the base flag once Qwen NextN / DeepSeek V3
+    # NextN also pre-slice.
+    pre_attention_trim: bool = True
+
     def __init__(
         self,
         config: LlamaConfig,