From 350a90b524bacdf670fd3fb62adb21822c255cc8 Mon Sep 17 00:00:00 2001 From: Abiba Date: Tue, 19 May 2026 21:11:34 +0000 Subject: [PATCH] fix: sync tier 4 default threshold to 50000 tokens (was stale at 4000) --- router/router.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/router/router.py b/router/router.py index fb2b6e1..0a47292 100644 --- a/router/router.py +++ b/router/router.py @@ -188,7 +188,7 @@ def route(rd, tier): if result: return result # TIER 4: Default — MoE first, VLM helps, Dense last (slow) - if t <= 4000: + if t <= 50000: candidates = [m for m in ["qwen3.6-35B-A3B","qwen3.5-9b-vlm","qwen3.6-27B-code"] if m in avail] result = select_best_gpu(candidates, "default") if result: return result