diff --git a/single_shot_inference.py b/single_shot_inference.py
index 105953ed..d456bda8 100644
--- a/single_shot_inference.py
+++ b/single_shot_inference.py
@@ -277,12 +277,15 @@ class HcHead:
 # Production FMHA
 # =====================================================================
 def _run_production_fmha(q_heads, all_kv, n_h, hd, T, seq_len, scale, dev, li, w, pfx):
-    from dsv4.kernels.attention.production import dsv4_attention_per_head
-    q = q_heads.permute(1, 0, 2).contiguous(); k = all_kv.unsqueeze(0).contiguous(); v = k.clone()
+    from dsv4.kernels.attention.production import dsv4_attention
+    # Head-packed dispatch: single kernel launch for all 128 heads (MQA: 1 KV head shared)
+    q = q_heads.permute(1, 0, 2).contiguous()  # (n_h, T, hd)
+    k = all_kv.unsqueeze(0).contiguous()          # (1, N, hd) — MQA single KV head
+    v = k.clone()
     sinks = w.get(f"{pfx}.sinks"); sink_bias = None
     if sinks is not None: sink_bias = sinks.to(device=dev).float().reshape(n_h)
-    attn_out = dsv4_attention_per_head(q=q, k=k, v=v, scale=scale, n_comp=0, sink_bias=sink_bias)
-    return attn_out.permute(1, 0, 2)
+    attn_out = dsv4_attention(q=q, k=k, v=v, scale=scale, n_comp=0, sink_bias=sink_bias)
+    return attn_out.permute(1, 0, 2)  # (T, n_h, hd)
 
 # =====================================================================
 # Attention — ALL production kernels