diff --git a/.modelhub_state/architecture_compatibility_blacklist.json b/.modelhub_state/architecture_compatibility_blacklist.json
index e30ad5e7ff..59804db2a4 100644
--- a/.modelhub_state/architecture_compatibility_blacklist.json
+++ b/.modelhub_state/architecture_compatibility_blacklist.json
@@ -3261,7 +3261,7 @@
       "taskType": "text-generation"
     }
   },
-  "generatedAt": "2026-09-27T13:38:02.660533+00:00",
+  "generatedAt": "2026-09-27T13:43:23.104832+00:00",
   "summary": {
     "activeBlockCount": 164,
     "byGpuFramework": {
diff --git a/.modelhub_state/market_intelligence.json b/.modelhub_state/market_intelligence.json
index 5006089a81..7f3a0fbbb4 100644
--- a/.modelhub_state/market_intelligence.json
+++ b/.modelhub_state/market_intelligence.json
@@ -1,8 +1,8 @@
 {
-  "communityAttemptedAt": "2026-09-27T13:27:20.547714+00:00",
+  "communityAttemptedAt": "2026-09-27T13:43:57.868379+00:00",
   "communityError": null,
   "communitySample": {},
-  "communityUpdatedAt": "2026-09-27T13:27:20.547714+00:00",
+  "communityUpdatedAt": "2026-09-27T13:43:57.868379+00:00",
   "frameworkAttemptedAt": "2026-09-27T11:27:14.277836+00:00",
   "frameworkError": null,
   "frameworkStats": {
@@ -396,7 +396,7 @@
     }
   },
   "frameworkUpdatedAt": null,
-  "generatedAt": "2026-09-27T13:42:17.478809+00:00",
+  "generatedAt": "2026-09-27T13:43:57.868379+00:00",
   "gpuStats": {
     "Ascend_910-b4": {
       "available": true,
diff --git a/.modelhub_state/official_capabilities.json b/.modelhub_state/official_capabilities.json
index fb186b43c1..7cd7af4103 100644
--- a/.modelhub_state/official_capabilities.json
+++ b/.modelhub_state/official_capabilities.json
@@ -1,5 +1,5 @@
 {
-  "catalogUpdatedAt": "2026-09-27T13:42:17.478809+00:00",
+  "catalogUpdatedAt": "2026-09-27T13:43:57.868379+00:00",
   "configuredTaskTypes": [
     "text-generation"
   ],
@@ -56,7 +56,7 @@
     "time-series-forecasting"
   ],
   "errors": [],
-  "generatedAt": "2026-09-27T13:42:21.100949+00:00",
+  "generatedAt": "2026-09-27T13:44:26.255835+00:00",
   "gpuCatalog": {
     "Ascend_910-b3": {
       "canVerify": true,
@@ -162,6 +162,14 @@
       ],
       "updatedAt": "2026-09-27T04:34:19.773251+00:00"
     },
+    "https://modelscope.cn/models/AXERA-TECH/Qwen3-1.7B-GPTQ-Int4|2026-09-23T17:58:48+00:00|Ascend_910-b4": {
+      "taskTypes": [
+        "text-generation",
+        "text-to-image-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:19.758279+00:00"
+    },
     "https://modelscope.cn/models/AXERA-TECH/Qwen3-1.7B-GPTQ-Int4|2026-09-23T17:58:48+00:00|Cambricon_mlu-370-x4": {
       "taskTypes": [
         "text-generation",
@@ -178,12 +186,34 @@
       ],
       "updatedAt": "2026-09-27T04:32:30.956535+00:00"
     },
+    "https://modelscope.cn/models/AXERA-TECH/Qwen3-1.7B-GPTQ-Int4|2026-09-23T17:58:48+00:00|Iluvatar_mrv-100": {
+      "taskTypes": [
+        "asr",
+        "text-generation",
+        "text-to-image-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:19.795005+00:00"
+    },
     "https://modelscope.cn/models/AXERA-TECH/Qwen3-1.7B-GPTQ-Int4|2026-09-23T17:58:48+00:00|Kunlunxin_p-800": {
       "taskTypes": [
         "text-generation"
       ],
       "updatedAt": "2026-09-27T04:36:00.838010+00:00"
     },
+    "https://modelscope.cn/models/AXERA-TECH/Qwen3-1.7B-GPTQ-Int4|2026-09-23T17:58:48+00:00|Mthreads_s4000": {
+      "taskTypes": [
+        "text-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:19.686967+00:00"
+    },
+    "https://modelscope.cn/models/AXERA-TECH/Qwen3-1.7B-GPTQ-Int4|2026-09-23T17:58:48+00:00|Sunrise_pt-200-x1": {
+      "taskTypes": [
+        "text-generation"
+      ],
+      "updatedAt": "2026-09-27T13:44:19.863662+00:00"
+    },
     "https://modelscope.cn/models/AXERA-TECH/Qwen3-1.7B-GPTQ-Int4|2026-09-23T17:58:48+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -258,12 +288,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.858315+00:00"
     },
-    "https://modelscope.cn/models/AlexWortega/llm-cipher-reasoning-loras|2026-09-03T05:28:45+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:56.163689+00:00"
-    },
     "https://modelscope.cn/models/AlexWortega/llm-cipher-reasoning-loras|2026-09-03T05:28:45+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -550,6 +574,12 @@
       ],
       "updatedAt": "2026-09-27T04:34:17.871507+00:00"
     },
+    "https://modelscope.cn/models/IFM/K2-Horizon-7B-GGUF|2026-09-23T16:46:11+00:00|Sunrise_pt-200-x1": {
+      "taskTypes": [
+        "text-generation"
+      ],
+      "updatedAt": "2026-09-27T13:44:19.953166+00:00"
+    },
     "https://modelscope.cn/models/Inferact/GLM-5.3-NVFP4|2026-08-28T19:40:14+00:00|Ascend_910-b4": {
       "taskTypes": [
         "text-generation",
@@ -747,6 +777,13 @@
       ],
       "updatedAt": "2026-09-27T04:34:18.118117+00:00"
     },
+    "https://modelscope.cn/models/JXW7777/TGAI_NB|2026-09-20T06:13:38+00:00|Mthreads_s4000": {
+      "taskTypes": [
+        "text-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:25.952154+00:00"
+    },
     "https://modelscope.cn/models/JXW7777/TGAI_NB|2026-09-20T06:13:38+00:00|Sunrise_pt-200-x1": {
       "taskTypes": [
         "text-generation"
@@ -1002,14 +1039,6 @@
       ],
       "updatedAt": "2026-09-27T12:03:12.662829+00:00"
     },
-    "https://modelscope.cn/models/OpenBMB/MiniCPM5-2B-DSpark-GGUF|2026-09-09T18:25:18+00:00|Cambricon_mlu-370-x4": {
-      "taskTypes": [
-        "text-generation",
-        "text-to-image-generation",
-        "visual-multi-modal"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.912354+00:00"
-    },
     "https://modelscope.cn/models/OpenBMB/MiniCPM5-2B-DSpark-GGUF|2026-09-09T18:25:18+00:00|Cambricon_mlu-370-x8": {
       "taskTypes": [
         "text-generation",
@@ -1054,12 +1083,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.661600+00:00"
     },
-    "https://modelscope.cn/models/OpenBMB/MiniCPM5-2B-DSpark-GGUF|2026-09-09T18:25:18+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.861625+00:00"
-    },
     "https://modelscope.cn/models/OpenBMB/MiniCPM5-2B-DSpark-GGUF|2026-09-09T18:25:18+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -1082,14 +1105,6 @@
       ],
       "updatedAt": "2026-09-27T12:03:12.739259+00:00"
     },
-    "https://modelscope.cn/models/OpenBMB/MiniCPM5-2B-DSpark|2026-09-07T18:16:04+00:00|Cambricon_mlu-370-x4": {
-      "taskTypes": [
-        "text-generation",
-        "text-to-image-generation",
-        "visual-multi-modal"
-      ],
-      "updatedAt": "2026-09-27T00:10:56.086837+00:00"
-    },
     "https://modelscope.cn/models/OpenBMB/MiniCPM5-2B-DSpark|2026-09-07T18:16:04+00:00|Cambricon_mlu-370-x8": {
       "taskTypes": [
         "text-generation",
@@ -1134,12 +1149,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.728855+00:00"
     },
-    "https://modelscope.cn/models/OpenBMB/MiniCPM5-2B-DSpark|2026-09-07T18:16:04+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:56.054176+00:00"
-    },
     "https://modelscope.cn/models/OpenBMB/MiniCPM5-2B-DSpark|2026-09-07T18:16:04+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -1234,6 +1243,44 @@
       ],
       "updatedAt": "2026-09-27T03:09:09.680006+00:00"
     },
+    "https://modelscope.cn/models/Qwen/Qwen3Guard-Stream-0.6B|2026-09-27T13:41:58+00:00|Iluvatar_mrv-100": {
+      "taskTypes": [
+        "asr",
+        "text-generation",
+        "text-to-image-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:17.257773+00:00"
+    },
+    "https://modelscope.cn/models/Qwen/Qwen3Guard-Stream-0.6B|2026-09-27T13:41:58+00:00|Kunlunxin_p-800": {
+      "taskTypes": [
+        "text-generation"
+      ],
+      "updatedAt": "2026-09-27T13:44:17.356597+00:00"
+    },
+    "https://modelscope.cn/models/Qwen/Qwen3Guard-Stream-0.6B|2026-09-27T13:41:58+00:00|MetaX_c-500": {
+      "taskTypes": [
+        "feature_emb",
+        "text-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:17.058559+00:00"
+    },
+    "https://modelscope.cn/models/Qwen/Qwen3Guard-Stream-0.6B|2026-09-27T13:41:58+00:00|Mthreads_s4000": {
+      "taskTypes": [
+        "text-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:16.957871+00:00"
+    },
+    "https://modelscope.cn/models/Qwen/Qwen3Guard-Stream-0.6B|2026-09-27T13:41:58+00:00|hygon_k100-ai": {
+      "taskTypes": [
+        "text-generation",
+        "text-to-image-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:17.155098+00:00"
+    },
     "https://modelscope.cn/models/SC117/Ornith-1.5-35B-A3B-Heretic-MTP-APEX-GGUF|2026-08-23T14:14:17+00:00|Kunlunxin_p-800": {
       "taskTypes": [
         "text-generation"
@@ -1260,14 +1307,6 @@
       ],
       "updatedAt": "2026-09-27T12:03:12.486255+00:00"
     },
-    "https://modelscope.cn/models/SKYEVAL/Three_Kingdoms_LLM_Arena|2026-09-14T09:18:14+00:00|Cambricon_mlu-370-x4": {
-      "taskTypes": [
-        "text-generation",
-        "text-to-image-generation",
-        "visual-multi-modal"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.778835+00:00"
-    },
     "https://modelscope.cn/models/SKYEVAL/Three_Kingdoms_LLM_Arena|2026-09-14T09:18:14+00:00|Cambricon_mlu-370-x8": {
       "taskTypes": [
         "text-generation",
@@ -1312,12 +1351,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.480034+00:00"
     },
-    "https://modelscope.cn/models/SKYEVAL/Three_Kingdoms_LLM_Arena|2026-09-14T09:18:14+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.673015+00:00"
-    },
     "https://modelscope.cn/models/SKYEVAL/Three_Kingdoms_LLM_Arena|2026-09-14T09:18:14+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -1504,12 +1537,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.667966+00:00"
     },
-    "https://modelscope.cn/models/TokenRhythm/NeoHorse-1-4B-GGUF|2026-09-09T03:26:23+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.963056+00:00"
-    },
     "https://modelscope.cn/models/TokenRhythm/NeoHorse-1-9B-GGUF|2026-09-09T03:27:14+00:00|Kunlunxin_p-800": {
       "taskTypes": [
         "text-generation"
@@ -1706,6 +1733,15 @@
       ],
       "updatedAt": "2026-09-27T04:32:31.362347+00:00"
     },
+    "https://modelscope.cn/models/XingChen-AGI/Xing4.0-29B-A4B-GGUF|2026-09-21T03:33:10+00:00|Iluvatar_mrv-100": {
+      "taskTypes": [
+        "asr",
+        "text-generation",
+        "text-to-image-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:21.698646+00:00"
+    },
     "https://modelscope.cn/models/XingChen-AGI/Xing4.0-29B-A4B-GGUF|2026-09-21T03:33:10+00:00|Kunlunxin_p-800": {
       "taskTypes": [
         "text-generation"
@@ -1727,6 +1763,12 @@
       ],
       "updatedAt": "2026-09-27T12:09:46.453156+00:00"
     },
+    "https://modelscope.cn/models/XingChen-AGI/Xing4.0-29B-A4B-GGUF|2026-09-21T03:33:10+00:00|Sunrise_pt-200-x1": {
+      "taskTypes": [
+        "text-generation"
+      ],
+      "updatedAt": "2026-09-27T13:44:21.770832+00:00"
+    },
     "https://modelscope.cn/models/XingChen-AGI/Xing4.0-29B-A4B-GGUF|2026-09-21T03:33:10+00:00|Vastai_va16": {
       "taskTypes": [
         "text-generation"
@@ -1821,14 +1863,6 @@
       ],
       "updatedAt": "2026-09-27T12:03:12.858526+00:00"
     },
-    "https://modelscope.cn/models/agentionai/Qwen3.8-27B-DFlash2-ROCmFP4-FAST-GGUF|2026-09-03T03:52:57+00:00|Cambricon_mlu-370-x4": {
-      "taskTypes": [
-        "text-generation",
-        "text-to-image-generation",
-        "visual-multi-modal"
-      ],
-      "updatedAt": "2026-09-27T00:10:56.275613+00:00"
-    },
     "https://modelscope.cn/models/agentionai/Qwen3.8-27B-DFlash2-ROCmFP4-FAST-GGUF|2026-09-03T03:52:57+00:00|Cambricon_mlu-370-x8": {
       "taskTypes": [
         "text-generation",
@@ -1873,12 +1907,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.859043+00:00"
     },
-    "https://modelscope.cn/models/agentionai/Qwen3.8-27B-DFlash2-ROCmFP4-FAST-GGUF|2026-09-03T03:52:57+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:56.164173+00:00"
-    },
     "https://modelscope.cn/models/agentionai/Qwen3.8-27B-DFlash2-ROCmFP4-FAST-GGUF|2026-09-03T03:52:57+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -2367,12 +2395,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.562422+00:00"
     },
-    "https://modelscope.cn/models/cix/DeepSeek-R1-Distill-Qwen-7B-GGUF|2026-09-11T07:54:02+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.858592+00:00"
-    },
     "https://modelscope.cn/models/dealignai/Bonsai-2-27B-1bit-CRACK-GGUF|2026-09-18T13:05:43+00:00|Ascend_910-b4": {
       "taskTypes": [
         "text-generation",
@@ -2387,14 +2409,6 @@
       ],
       "updatedAt": "2026-09-27T12:03:12.396561+00:00"
     },
-    "https://modelscope.cn/models/dealignai/Bonsai-2-27B-1bit-CRACK-GGUF|2026-09-18T13:05:43+00:00|Cambricon_mlu-370-x4": {
-      "taskTypes": [
-        "text-generation",
-        "text-to-image-generation",
-        "visual-multi-modal"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.578635+00:00"
-    },
     "https://modelscope.cn/models/dealignai/Bonsai-2-27B-1bit-CRACK-GGUF|2026-09-18T13:05:43+00:00|Cambricon_mlu-370-x8": {
       "taskTypes": [
         "text-generation",
@@ -2410,7 +2424,7 @@
         "text-to-image-generation",
         "visual-multi-modal"
       ],
-      "updatedAt": "2026-09-27T00:10:55.377054+00:00"
+      "updatedAt": "2026-09-27T13:44:26.255835+00:00"
     },
     "https://modelscope.cn/models/dealignai/Bonsai-2-27B-1bit-CRACK-GGUF|2026-09-18T13:05:43+00:00|Kunlunxin_p-800": {
       "taskTypes": [
@@ -2431,7 +2445,7 @@
         "text-generation",
         "visual-multi-modal"
       ],
-      "updatedAt": "2026-09-27T00:10:55.269693+00:00"
+      "updatedAt": "2026-09-27T13:44:26.191406+00:00"
     },
     "https://modelscope.cn/models/dealignai/Bonsai-2-27B-1bit-CRACK-GGUF|2026-09-18T13:05:43+00:00|Sunrise_pt-200-x1": {
       "taskTypes": [
@@ -2439,12 +2453,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.379345+00:00"
     },
-    "https://modelscope.cn/models/dealignai/Bonsai-2-27B-1bit-CRACK-GGUF|2026-09-18T13:05:43+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.477840+00:00"
-    },
     "https://modelscope.cn/models/dealignai/Bonsai-2-27B-1bit-CRACK-GGUF|2026-09-18T13:05:43+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -2467,14 +2475,6 @@
       ],
       "updatedAt": "2026-09-27T12:03:12.266743+00:00"
     },
-    "https://modelscope.cn/models/dealignai/Bonsai-2-27B-Ternary-CRACK-GGUF|2026-09-18T14:52:28+00:00|Cambricon_mlu-370-x4": {
-      "taskTypes": [
-        "text-generation",
-        "text-to-image-generation",
-        "visual-multi-modal"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.470792+00:00"
-    },
     "https://modelscope.cn/models/dealignai/Bonsai-2-27B-Ternary-CRACK-GGUF|2026-09-18T14:52:28+00:00|Cambricon_mlu-370-x8": {
       "taskTypes": [
         "text-generation",
@@ -2506,18 +2506,19 @@
       ],
       "updatedAt": "2026-09-27T04:34:18.357212+00:00"
     },
+    "https://modelscope.cn/models/dealignai/Bonsai-2-27B-Ternary-CRACK-GGUF|2026-09-18T14:52:28+00:00|Mthreads_s4000": {
+      "taskTypes": [
+        "text-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:26.095626+00:00"
+    },
     "https://modelscope.cn/models/dealignai/Bonsai-2-27B-Ternary-CRACK-GGUF|2026-09-18T14:52:28+00:00|Sunrise_pt-200-x1": {
       "taskTypes": [
         "text-generation"
       ],
       "updatedAt": "2026-09-27T13:42:19.302352+00:00"
     },
-    "https://modelscope.cn/models/dealignai/Bonsai-2-27B-Ternary-CRACK-GGUF|2026-09-18T14:52:28+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.377397+00:00"
-    },
     "https://modelscope.cn/models/dealignai/Bonsai-2-27B-Ternary-CRACK-GGUF|2026-09-18T14:52:28+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -2939,14 +2940,6 @@
       ],
       "updatedAt": "2026-09-27T12:03:12.473620+00:00"
     },
-    "https://modelscope.cn/models/icychick/Qwen3.5-text-0.8B-GGUF|2026-09-14T05:41:13+00:00|Cambricon_mlu-370-x4": {
-      "taskTypes": [
-        "text-generation",
-        "text-to-image-generation",
-        "visual-multi-modal"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.761356+00:00"
-    },
     "https://modelscope.cn/models/icychick/Qwen3.5-text-0.8B-GGUF|2026-09-14T05:41:13+00:00|Cambricon_mlu-370-x8": {
       "taskTypes": [
         "text-generation",
@@ -2991,12 +2984,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.470134+00:00"
     },
-    "https://modelscope.cn/models/icychick/Qwen3.5-text-0.8B-GGUF|2026-09-14T05:41:13+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.673451+00:00"
-    },
     "https://modelscope.cn/models/icychick/Qwen3.5-text-0.8B-GGUF|2026-09-14T05:41:13+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -3357,6 +3344,13 @@
       ],
       "updatedAt": "2026-09-27T04:34:18.117784+00:00"
     },
+    "https://modelscope.cn/models/iyangqing/st-image|2026-09-20T08:37:04+00:00|Mthreads_s4000": {
+      "taskTypes": [
+        "text-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:25.950534+00:00"
+    },
     "https://modelscope.cn/models/iyangqing/st-image|2026-09-20T08:37:04+00:00|Sunrise_pt-200-x1": {
       "taskTypes": [
         "text-generation"
@@ -3668,14 +3662,6 @@
       ],
       "updatedAt": "2026-09-27T12:03:12.666119+00:00"
     },
-    "https://modelscope.cn/models/laion/GLM-4.6-stackexchange-overflow-sandboxes-32eps-65k-reasoning_global-batch-size_32_Qwen3-32B|2026-09-09T17:06:17+00:00|Cambricon_mlu-370-x4": {
-      "taskTypes": [
-        "text-generation",
-        "text-to-image-generation",
-        "visual-multi-modal"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.912408+00:00"
-    },
     "https://modelscope.cn/models/laion/GLM-4.6-stackexchange-overflow-sandboxes-32eps-65k-reasoning_global-batch-size_32_Qwen3-32B|2026-09-09T17:06:17+00:00|Cambricon_mlu-370-x8": {
       "taskTypes": [
         "text-generation",
@@ -3720,12 +3706,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.664637+00:00"
     },
-    "https://modelscope.cn/models/laion/GLM-4.6-stackexchange-overflow-sandboxes-32eps-65k-reasoning_global-batch-size_32_Qwen3-32B|2026-09-09T17:06:17+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.861893+00:00"
-    },
     "https://modelscope.cn/models/laion/GLM-4.6-stackexchange-overflow-sandboxes-32eps-65k-reasoning_global-batch-size_32_Qwen3-32B|2026-09-09T17:06:17+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -3867,6 +3847,22 @@
       ],
       "updatedAt": "2026-09-27T03:15:47.463281+00:00"
     },
+    "https://modelscope.cn/models/mlx-community/LFM2.5-1.2B-Instruct-6bit|2026-09-23T17:53:19+00:00|Iluvatar_mrv-100": {
+      "taskTypes": [
+        "asr",
+        "text-generation",
+        "text-to-image-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:19.758244+00:00"
+    },
+    "https://modelscope.cn/models/mlx-community/LFM2.5-1.2B-Instruct-6bit|2026-09-23T17:53:19+00:00|Mthreads_s4000": {
+      "taskTypes": [
+        "text-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:19.708497+00:00"
+    },
     "https://modelscope.cn/models/mlx-community/LFM2.5-1.2B-Instruct-6bit|2026-09-23T17:53:19+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -4136,7 +4132,7 @@
         "text-generation",
         "visual-multi-modal"
       ],
-      "updatedAt": "2026-09-27T00:10:55.268965+00:00"
+      "updatedAt": "2026-09-27T13:44:26.153695+00:00"
     },
     "https://modelscope.cn/models/nv-community/Qwen3.8-27B-NVFP4|2026-09-18T13:36:05+00:00|hygon_k100-ai": {
       "taskTypes": [
@@ -4270,14 +4266,6 @@
       ],
       "updatedAt": "2026-09-27T12:03:12.396700+00:00"
     },
-    "https://modelscope.cn/models/prism-ml/Ternary-Bonsai-2-27B-gguf-dev|2026-09-18T05:30:35+00:00|Cambricon_mlu-370-x4": {
-      "taskTypes": [
-        "text-generation",
-        "text-to-image-generation",
-        "visual-multi-modal"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.578516+00:00"
-    },
     "https://modelscope.cn/models/prism-ml/Ternary-Bonsai-2-27B-gguf-dev|2026-09-18T05:30:35+00:00|Cambricon_mlu-370-x8": {
       "taskTypes": [
         "text-generation",
@@ -4314,7 +4302,7 @@
         "text-generation",
         "visual-multi-modal"
       ],
-      "updatedAt": "2026-09-27T00:10:55.377131+00:00"
+      "updatedAt": "2026-09-27T13:44:26.251725+00:00"
     },
     "https://modelscope.cn/models/prism-ml/Ternary-Bonsai-2-27B-gguf-dev|2026-09-18T05:30:35+00:00|Sunrise_pt-200-x1": {
       "taskTypes": [
@@ -4322,12 +4310,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.379470+00:00"
     },
-    "https://modelscope.cn/models/prism-ml/Ternary-Bonsai-2-27B-gguf-dev|2026-09-18T05:30:35+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.477889+00:00"
-    },
     "https://modelscope.cn/models/prism-ml/Ternary-Bonsai-2-27B-gguf-dev|2026-09-18T05:30:35+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -4354,12 +4336,29 @@
       ],
       "updatedAt": "2026-09-27T04:34:19.363239+00:00"
     },
+    "https://modelscope.cn/models/prithivMLmods/APUS-OpenJev-v1-4B-GGUF|2026-09-22T18:59:12+00:00|Ascend_910-b4": {
+      "taskTypes": [
+        "text-generation",
+        "text-to-image-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:20.791241+00:00"
+    },
     "https://modelscope.cn/models/prithivMLmods/APUS-OpenJev-v1-4B-GGUF|2026-09-22T18:59:12+00:00|Biren_166m": {
       "taskTypes": [
         "text-generation"
       ],
       "updatedAt": "2026-09-27T12:03:12.084894+00:00"
     },
+    "https://modelscope.cn/models/prithivMLmods/APUS-OpenJev-v1-4B-GGUF|2026-09-22T18:59:12+00:00|Iluvatar_mrv-100": {
+      "taskTypes": [
+        "asr",
+        "text-generation",
+        "text-to-image-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:20.860961+00:00"
+    },
     "https://modelscope.cn/models/prithivMLmods/APUS-OpenJev-v1-4B-GGUF|2026-09-22T18:59:12+00:00|Kunlunxin_p-800": {
       "taskTypes": [
         "text-generation"
@@ -4374,6 +4373,19 @@
       ],
       "updatedAt": "2026-09-27T04:34:17.987883+00:00"
     },
+    "https://modelscope.cn/models/prithivMLmods/APUS-OpenJev-v1-4B-GGUF|2026-09-22T18:59:12+00:00|Mthreads_s4000": {
+      "taskTypes": [
+        "text-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:20.750897+00:00"
+    },
+    "https://modelscope.cn/models/prithivMLmods/APUS-OpenJev-v1-4B-GGUF|2026-09-22T18:59:12+00:00|Sunrise_pt-200-x1": {
+      "taskTypes": [
+        "text-generation"
+      ],
+      "updatedAt": "2026-09-27T13:44:20.960268+00:00"
+    },
     "https://modelscope.cn/models/prithivMLmods/APUS-OpenJev-v1-4B-GGUF|2026-09-22T18:59:12+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -4516,6 +4528,12 @@
       ],
       "updatedAt": "2026-09-27T04:34:17.988958+00:00"
     },
+    "https://modelscope.cn/models/prithivMLmods/Orion-9B-Agentic-CodeCore-Merge-GGUF|2026-09-25T18:54:04+00:00|Kunlunxin_p-800": {
+      "taskTypes": [
+        "text-generation"
+      ],
+      "updatedAt": "2026-09-27T13:44:18.348643+00:00"
+    },
     "https://modelscope.cn/models/prithivMLmods/Qwen-Image-2.1-PE-T2I-GGUF|2026-09-20T18:32:08+00:00|Biren_166m": {
       "taskTypes": [
         "text-generation"
@@ -4614,6 +4632,12 @@
       ],
       "updatedAt": "2026-09-27T01:58:32.025413+00:00"
     },
+    "https://modelscope.cn/models/prithivMLmods/blink-mimo-9b-GGUF|2026-09-25T18:56:15+00:00|Kunlunxin_p-800": {
+      "taskTypes": [
+        "text-generation"
+      ],
+      "updatedAt": "2026-09-27T13:44:18.350078+00:00"
+    },
     "https://modelscope.cn/models/shensanshu/ninfer-ada-ternary|2026-09-21T03:08:27+00:00|Ascend_910-b4": {
       "taskTypes": [
         "text-generation",
@@ -4746,12 +4770,6 @@
       ],
       "updatedAt": "2026-09-27T13:42:19.775295+00:00"
     },
-    "https://modelscope.cn/models/svvkii/qwen2.5-7b-5persona-lora|2026-09-03T09:05:03+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:56.162844+00:00"
-    },
     "https://modelscope.cn/models/svvkii/qwen2.5-7b-5persona-lora|2026-09-03T09:05:03+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -5265,14 +5283,6 @@
       ],
       "updatedAt": "2026-09-27T12:03:12.266194+00:00"
     },
-    "https://modelscope.cn/models/wufubaoa/Nlp_Polylm_Assistant_13B_Text_Generation-15B-GGUF|2026-09-19T02:53:32+00:00|Cambricon_mlu-370-x4": {
-      "taskTypes": [
-        "text-generation",
-        "text-to-image-generation",
-        "visual-multi-modal"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.578609+00:00"
-    },
     "https://modelscope.cn/models/wufubaoa/Nlp_Polylm_Assistant_13B_Text_Generation-15B-GGUF|2026-09-19T02:53:32+00:00|Cambricon_mlu-370-x8": {
       "taskTypes": [
         "text-generation",
@@ -5288,7 +5298,7 @@
         "text-to-image-generation",
         "visual-multi-modal"
       ],
-      "updatedAt": "2026-09-27T00:10:55.376996+00:00"
+      "updatedAt": "2026-09-27T13:44:26.098924+00:00"
     },
     "https://modelscope.cn/models/wufubaoa/Nlp_Polylm_Assistant_13B_Text_Generation-15B-GGUF|2026-09-19T02:53:32+00:00|Kunlunxin_p-800": {
       "taskTypes": [
@@ -5304,18 +5314,19 @@
       ],
       "updatedAt": "2026-09-27T04:34:18.356901+00:00"
     },
+    "https://modelscope.cn/models/wufubaoa/Nlp_Polylm_Assistant_13B_Text_Generation-15B-GGUF|2026-09-19T02:53:32+00:00|Mthreads_s4000": {
+      "taskTypes": [
+        "text-generation",
+        "visual-multi-modal"
+      ],
+      "updatedAt": "2026-09-27T13:44:26.051496+00:00"
+    },
     "https://modelscope.cn/models/wufubaoa/Nlp_Polylm_Assistant_13B_Text_Generation-15B-GGUF|2026-09-19T02:53:32+00:00|Sunrise_pt-200-x1": {
       "taskTypes": [
         "text-generation"
       ],
       "updatedAt": "2026-09-27T13:42:19.301786+00:00"
     },
-    "https://modelscope.cn/models/wufubaoa/Nlp_Polylm_Assistant_13B_Text_Generation-15B-GGUF|2026-09-19T02:53:32+00:00|Vastai_va16": {
-      "taskTypes": [
-        "text-generation"
-      ],
-      "updatedAt": "2026-09-27T00:10:55.452565+00:00"
-    },
     "https://modelscope.cn/models/wufubaoa/Nlp_Polylm_Assistant_13B_Text_Generation-15B-GGUF|2026-09-19T02:53:32+00:00|hygon_k100-ai": {
       "taskTypes": [
         "text-generation",
@@ -6186,6 +6197,6 @@
       "updateTime": "2025-12-22 08:59:53"
     }
   ],
-  "taskTreeUpdatedAt": "2026-09-27T13:42:17.478809+00:00",
+  "taskTreeUpdatedAt": "2026-09-27T13:43:57.868379+00:00",
   "version": 1
 }
diff --git a/.modelhub_state/outcome_checkpoint.json b/.modelhub_state/outcome_checkpoint.json
index 7fcbeeefd2..8ea98d6940 100644
--- a/.modelhub_state/outcome_checkpoint.json
+++ b/.modelhub_state/outcome_checkpoint.json
@@ -1,6 +1,6 @@
 {
-  "generatedAt": "2026-09-27T12:27:45.561155+00:00",
-  "lastSyncTime": "2026-09-27T12:27:45.263970+00:00",
+  "generatedAt": "2026-09-27T13:43:23.022567+00:00",
+  "lastSyncTime": "2026-09-27T13:43:22.766786+00:00",
   "recentLimit": 300,
   "report": {
     "architectureCompatibilityBlocks": {
@@ -3845,25 +3845,25 @@
         "decisionSuccessRate": 0.3636,
         "decisionTotal": 22,
         "failureBreakdown": {
-          "ambiguous_runtime": 39,
+          "ambiguous_runtime": 40,
           "framework_architecture_unsupported": 11,
           "model_load": 3,
           "platform_infrastructure": 1,
           "参数/模板问题": 1,
           "验证失败": 3
         },
-        "failureCount": 58,
-        "failureRate": 0.8788,
+        "failureCount": 59,
+        "failureRate": 0.8806,
         "framework": "vllm-mlu",
         "pendingCount": 0,
         "pendingRate": 0.0,
         "platformFailureCount": 1,
         "successCount": 8,
-        "successRate": 0.1212,
+        "successRate": 0.1194,
         "targetGpu": "Cambricon_mlu-370-x4",
         "taskType": "text-generation",
-        "total": 66,
-        "unresolvedFailureCount": 43
+        "total": 67,
+        "unresolvedFailureCount": 44
       },
       "Cambricon_mlu-370-x4|vllm|text-generation": {
         "attributableFailureCount": 24,
@@ -6015,7 +6015,7 @@
         "decisionSuccessRate": 0.4167,
         "decisionTotal": 48,
         "failureBreakdown": {
-          "ambiguous_runtime": 86,
+          "ambiguous_runtime": 87,
           "framework_architecture_unsupported": 22,
           "model_load": 5,
           "platform_infrastructure": 1,
@@ -6023,15 +6023,15 @@
           "参数/模板问题": 4,
           "验证失败": 3
         },
-        "failureCount": 122,
-        "failureRate": 0.8592,
+        "failureCount": 123,
+        "failureRate": 0.8601,
         "pendingCount": 0,
         "pendingRate": 0.0,
         "platformFailureCount": 1,
         "successCount": 20,
-        "successRate": 0.1408,
-        "total": 142,
-        "unresolvedFailureCount": 93
+        "successRate": 0.1399,
+        "total": 143,
+        "unresolvedFailureCount": 94
       },
       "vllm-patch-tokenizer": {
         "attributableFailureCount": 55,
@@ -6134,7 +6134,7 @@
         "unresolvedFailureCount": 197
       }
     },
-    "generatedAt": "2026-09-27T12:27:45.548175+00:00",
+    "generatedAt": "2026-09-27T13:43:23.009548+00:00",
     "gpuSummaries": {
       "Ascend_910-b3": {
         "attributableFailureCount": 102,
@@ -6226,7 +6226,7 @@
         "decisionSuccessRate": 0.0894,
         "decisionTotal": 817,
         "failureBreakdown": {
-          "ambiguous_runtime": 499,
+          "ambiguous_runtime": 500,
           "architecture_compatibility": 53,
           "context_length": 48,
           "framework_architecture_unsupported": 297,
@@ -6239,15 +6239,15 @@
           "日志缺失": 13,
           "验证失败": 42
         },
-        "failureCount": 1601,
+        "failureCount": 1602,
         "failureRate": 0.9564,
         "pendingCount": 0,
         "pendingRate": 0.0,
         "platformFailureCount": 16,
         "successCount": 73,
         "successRate": 0.0436,
-        "total": 1674,
-        "unresolvedFailureCount": 841
+        "total": 1675,
+        "unresolvedFailureCount": 842
       },
       "Cambricon_mlu-370-x8": {
         "attributableFailureCount": 60,
@@ -10852,6 +10852,29 @@
         "total": 1,
         "unresolvedFailureCount": 1
       },
+      "Cambricon_mlu-370-x4|vllm-mlu|text-generation|mistral|compressed-tensors": {
+        "attributableFailureCount": 0,
+        "decisionFailureRate": 0.0,
+        "decisionSuccessRate": 0.0,
+        "decisionTotal": 0,
+        "failureBreakdown": {
+          "ambiguous_runtime": 1
+        },
+        "failureCount": 1,
+        "failureRate": 1.0,
+        "framework": "vllm-mlu",
+        "modelType": "mistral",
+        "pendingCount": 0,
+        "pendingRate": 0.0,
+        "platformFailureCount": 0,
+        "quantizationMethod": "compressed-tensors",
+        "successCount": 0,
+        "successRate": 0.0,
+        "targetGpu": "Cambricon_mlu-370-x4",
+        "taskType": "text-generation",
+        "total": 1,
+        "unresolvedFailureCount": 1
+      },
       "Cambricon_mlu-370-x4|vllm-mlu|text-generation|mpt|none": {
         "attributableFailureCount": 0,
         "decisionFailureRate": 0.0,
@@ -22912,10 +22935,10 @@
         "decisionSuccessRate": 0.0,
         "decisionTotal": 0,
         "failureBreakdown": {
-          "ambiguous_runtime": 4,
+          "ambiguous_runtime": 5,
           "参数/模板问题": 1
         },
-        "failureCount": 5,
+        "failureCount": 6,
         "failureRate": 1.0,
         "framework": "vllm-mlu",
         "lastPlatformFailureAt": null,
@@ -22927,8 +22950,8 @@
         "successRate": 0.0,
         "targetGpu": "Cambricon_mlu-370-x4",
         "taskType": "text-generation",
-        "total": 5,
-        "unresolvedFailureCount": 5
+        "total": 6,
+        "unresolvedFailureCount": 6
       },
       "Cambricon_mlu-370-x4|vllm|text-generation": {
         "attributableFailureCount": 1,
@@ -22938,10 +22961,10 @@
         "decisionSuccessRate": 0.0,
         "decisionTotal": 1,
         "failureBreakdown": {
-          "ambiguous_runtime": 4,
+          "ambiguous_runtime": 3,
           "framework_architecture_unsupported": 1
         },
-        "failureCount": 5,
+        "failureCount": 4,
         "failureRate": 1.0,
         "framework": "vllm",
         "lastPlatformFailureAt": null,
@@ -22953,8 +22976,8 @@
         "successRate": 0.0,
         "targetGpu": "Cambricon_mlu-370-x4",
         "taskType": "text-generation",
-        "total": 5,
-        "unresolvedFailureCount": 4
+        "total": 4,
+        "unresolvedFailureCount": 3
       },
       "Cambricon_mlu-370-x8|unknown|text-generation": {
         "attributableFailureCount": 0,
@@ -23957,6 +23980,31 @@
         "total": 3,
         "unresolvedFailureCount": 3
       },
+      "Cambricon_mlu-370-x4|vllm-mlu|text-generation|mistral|compressed-tensors": {
+        "attributableFailureCount": 0,
+        "consecutiveFailures": 0,
+        "decisionFailureRate": 0.0,
+        "decisionSuccessRate": 0.0,
+        "decisionTotal": 0,
+        "failureBreakdown": {
+          "ambiguous_runtime": 1
+        },
+        "failureCount": 1,
+        "failureRate": 1.0,
+        "framework": "vllm-mlu",
+        "lastTerminalAt": "2026-09-27T13:43:22.766786+00:00",
+        "modelType": "mistral",
+        "pendingCount": 0,
+        "pendingRate": 0.0,
+        "platformFailureCount": 0,
+        "quantizationMethod": "compressed-tensors",
+        "successCount": 0,
+        "successRate": 0.0,
+        "targetGpu": "Cambricon_mlu-370-x4",
+        "taskType": "text-generation",
+        "total": 1,
+        "unresolvedFailureCount": 1
+      },
       "Cambricon_mlu-370-x4|vllm-mlu|text-generation|starcoder2|compressed-tensors": {
         "attributableFailureCount": 0,
         "consecutiveFailures": 0,
@@ -23982,31 +24030,6 @@
         "total": 1,
         "unresolvedFailureCount": 1
       },
-      "Cambricon_mlu-370-x4|vllm|text-generation|gemma2|compressed-tensors": {
-        "attributableFailureCount": 0,
-        "consecutiveFailures": 0,
-        "decisionFailureRate": 0.0,
-        "decisionSuccessRate": 0.0,
-        "decisionTotal": 0,
-        "failureBreakdown": {
-          "ambiguous_runtime": 1
-        },
-        "failureCount": 1,
-        "failureRate": 1.0,
-        "framework": "vllm",
-        "lastTerminalAt": "2026-09-25T09:35:33.057138+00:00",
-        "modelType": "gemma2",
-        "pendingCount": 0,
-        "pendingRate": 0.0,
-        "platformFailureCount": 0,
-        "quantizationMethod": "compressed-tensors",
-        "successCount": 0,
-        "successRate": 0.0,
-        "targetGpu": "Cambricon_mlu-370-x4",
-        "taskType": "text-generation",
-        "total": 1,
-        "unresolvedFailureCount": 1
-      },
       "Cambricon_mlu-370-x4|vllm|text-generation|lfm2|none": {
         "attributableFailureCount": 1,
         "consecutiveFailures": 1,
@@ -33354,6 +33377,30 @@
         "total": 1,
         "unresolvedFailureCount": 1
       },
+      "Cambricon_mlu-370-x4|vllm-mlu|text-generation|mistral|compressed-tensors|33": {
+        "attributableFailureCount": 0,
+        "decisionFailureRate": 0.0,
+        "decisionSuccessRate": 0.0,
+        "decisionTotal": 0,
+        "failureBreakdown": {
+          "ambiguous_runtime": 1
+        },
+        "failureCount": 1,
+        "failureRate": 1.0,
+        "framework": "vllm-mlu",
+        "loadSizeLog2Bucket": 33,
+        "modelType": "mistral",
+        "pendingCount": 0,
+        "pendingRate": 0.0,
+        "platformFailureCount": 0,
+        "quantizationMethod": "compressed-tensors",
+        "successCount": 0,
+        "successRate": 0.0,
+        "targetGpu": "Cambricon_mlu-370-x4",
+        "taskType": "text-generation",
+        "total": 1,
+        "unresolvedFailureCount": 1
+      },
       "Cambricon_mlu-370-x4|vllm-mlu|text-generation|mpt|none|33": {
         "attributableFailureCount": 0,
         "decisionFailureRate": 0.0,
@@ -51314,15 +51361,15 @@
         "unresolvedFailureCount": 0
       }
     },
-    "terminalRecords": 17061,
-    "totalRecords": 17282,
+    "terminalRecords": 17062,
+    "totalRecords": 17283,
     "totals": {
       "attributableFailureCount": 6114,
       "decisionFailureRate": 0.8636,
       "decisionSuccessRate": 0.1364,
       "decisionTotal": 7080,
       "failureBreakdown": {
-        "ambiguous_runtime": 4420,
+        "ambiguous_runtime": 4421,
         "architecture_compatibility": 212,
         "attention_backend": 3,
         "backend_operator": 128,
@@ -51338,15 +51385,15 @@
         "日志缺失": 719,
         "验证失败": 676
       },
-      "failureCount": 16095,
+      "failureCount": 16096,
       "failureRate": 0.9434,
       "pendingCount": 0,
       "pendingRate": 0.0,
       "platformFailureCount": 938,
       "successCount": 966,
       "successRate": 0.0566,
-      "total": 17061,
-      "unresolvedFailureCount": 9043
+      "total": 17062,
+      "unresolvedFailureCount": 9044
     },
     "warnings": [
       "GPU Iluvatar_mrv-100 本地统计失败率偏高（≥50%），建议重点关注。",
@@ -51354,15 +51401,15 @@
       "GPU Ascend_910-b4 本地统计失败率偏高（≥50%），建议重点关注。",
       "GPU Cambricon_mlu-370-x4 本地统计失败率偏高（≥50%），建议重点关注。",
       "GPU MetaX_c-500 本地统计失败率偏高（≥50%），建议重点关注。",
-      "GPU Mthreads_s4000 本地统计失败率偏高（≥50%），建议重点关注。",
       "GPU Vastai_va16 本地统计失败率偏高（≥50%），建议重点关注。",
+      "GPU Iluvatar_bi-150 本地统计失败率偏高（≥50%），建议重点关注。",
       "GPU Ascend_910-b3 本地统计失败率偏高（≥50%），建议重点关注。",
       "GPU Sunrise_pt-200-x1 本地统计失败率偏高（≥50%），建议重点关注。",
       "GPU Biren_166m 本地统计失败率偏高（≥50%），建议重点关注。",
       "GPU hygon_k100-ai 本地统计失败率偏高（≥50%），建议重点关注。",
       "GPU Iluvatar_bi-100 本地统计失败率偏高（≥50%），建议重点关注。",
       "GPU Kunlunxin_p-800 本地统计失败率偏高（≥50%），建议重点关注。",
-      "GPU Iluvatar_bi-150 本地统计失败率偏高（≥50%），建议重点关注。",
+      "GPU Mthreads_s4000 本地统计失败率偏高（≥50%），建议重点关注。",
       "组合 Ascend_910-b3|vllm_tokenizer_patch|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
       "组合 Ascend_910-b3|vllm|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
       "组合 Ascend_910-b4|llamacpp|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
@@ -51386,7 +51433,6 @@
       "组合 Cambricon_mlu-370-x4|vllm|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
       "组合 Iluvatar_bi-150|transformers|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
       "组合 hygon_k100-ai|vllm-patch-tokenizer|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
-      "组合 Vastai_va16|vllm_fix_tokenizer|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
       "组合 Iluvatar_mrv-100|unknown|visual-multi-modal 近期失败集中，建议降低该 GPU+框架的提交优先级。",
       "组合 Mthreads_s4000|llamacpp|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
       "组合 Iluvatar_bi-100|vllm|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
@@ -51400,10 +51446,11 @@
       "组合 Cambricon_mlu-370-x4|unknown|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
       "组合 Mthreads_s4000|vllm|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
       "组合 Kunlunxin_p-800|vllm_tokenizer_patch|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
-      "组合 Iluvatar_bi-150|llamacpp|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。"
+      "组合 Iluvatar_bi-150|llamacpp|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。",
+      "组合 Vastai_va16|vllm_fix_tokenizer|text-generation 近期失败集中，建议降低该 GPU+框架的提交优先级。"
     ]
   },
   "storageMode": "decision_state_only",
-  "summarizedRecords": 17282,
+  "summarizedRecords": 17283,
   "version": 1
 }
diff --git a/.modelhub_state/recent_outcomes.jsonl b/.modelhub_state/recent_outcomes.jsonl
index d9ae4c6bf8..ce82b19d2d 100644
--- a/.modelhub_state/recent_outcomes.jsonl
+++ b/.modelhub_state/recent_outcomes.jsonl
@@ -201,6 +201,7 @@
 {"failReason": "参数/模板问题", "framework": "", "lastSyncTime": "2026-09-21T09:54:21.980336+00:00", "modelId": "nv-community/NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP4", "modelProfile": {}, "outcome": "failed", "status": "failed", "submitTime": "2026-09-21T09:35:50+00:00", "targetGpu": "Cambricon_mlu-370-x4", "taskId": "4969372", "taskType": "text-generation", "verifyResult": null}
 {"failReason": "ambiguous_runtime", "failureAction": "send_compact_profile_and_root_exception_to_llm", "failureCategory": "ambiguous_runtime", "failureClassificationReason": "execute_empty_result", "failureCode": "EXECUTE_EMPTY_RESULT", "failureDeterministic": false, "failureEnrichmentAttempts": 1, "failureEnrichmentError": null, "failureNeedsLlm": true, "failureScope": "unknown", "framework": "vllm-customized", "lastSyncTime": "2026-09-22T21:50:30.559334+00:00", "modelId": "solidrust/Llama-3-16B-Instruct-v0.1-AWQ", "modelProfile": {"architectures": ["LlamaForCausalLM"], "configFingerprint": "1b41437401598cde7836ccc25883635c2738b06a5af924b7497b6e747e655b4b", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 10261168632, "estimatedRequiredGiB": 11.478, "gpuMemoryEvidence": {"memoryGiB": 24.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 24.0, "maximumRepositorySizeGiB": 20.0, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "llama", "modelscopeFileSize": 10270608122, "modelscopeLicense": "other", "modelscopeParams": 16754741248, "modelscopeTags": ["license:other", "model_type:llama", "library:safetensors", "library:pytorch", "task:text-generation", "custom_tag:4-bit", "custom_tag:AWQ", "custom_tag:text-generation", "custom_tag:autotrain_compatible", "custom_tag:endpoints_compatible - mergekit - merge - facebook - meta - pytorch - llama - llama-3"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "awq", "repositoryOnDiskBytes": 10270608122}, "outcome": "failed", "status": "success", "submitTime": "2026-09-21T09:33:26.659264+00:00", "targetGpu": "Cambricon_mlu-370-x8", "taskId": "4998167", "taskType": "text-generation", "verifyResult": -1}
 {"failReason": "ambiguous_runtime", "failureAction": "send_compact_profile_and_root_exception_to_llm", "failureCategory": "ambiguous_runtime", "failureClassificationReason": "execute_empty_result", "failureCode": "EXECUTE_EMPTY_RESULT", "failureDeterministic": false, "failureEnrichmentAttempts": 1, "failureEnrichmentError": null, "failureNeedsLlm": true, "failureScope": "unknown", "framework": "vllm_fix_tokenizer", "lastSyncTime": "2026-09-23T10:48:31.671315+00:00", "modelId": "nm-testing/tinyllama-oneshot-w8a8-dynamic-token-v2", "modelProfile": {"architectures": ["LlamaForCausalLM"], "configFingerprint": "d8df464812ff2fd3c46dc89b0fc6a96370e694fab8e52179f1ada2cdd0ea92b9", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 1231252716, "estimatedRequiredGiB": 1.378, "gpuMemoryEvidence": {"memoryGiB": 100.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 100.0, "maximumRepositorySizeGiB": 83.333, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "llama", "modelscopeFileSize": 1233101396, "modelscopeLicense": null, "modelscopeParams": 1100048384, "modelscopeTags": ["model_type:llama", "library:safetensors", "library:pytorch", "task:text-generation"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "compressed-tensors", "recentProfileFeedback": {"attributableFailureCount": 0, "consecutiveFailures": 0, "decisionFailureRate": 0.0, "decisionSuccessRate": 0.0, "decisionTotal": 0, "failureBreakdown": {"ambiguous_runtime": 1}, "failureCount": 1, "failureRate": 1.0, "framework": "vllm_fix_tokenizer", "lastTerminalAt": "2026-09-21T08:44:09.587569+00:00", "modelType": "llama", "pendingCount": 0, "pendingRate": 0.0, "platformFailureCount": 0, "quantizationMethod": "compressed-tensors", "successCount": 0, "successRate": 0.0, "targetGpu": "Kunlunxin_p-800", "taskType": "text-generation", "total": 1, "unresolvedFailureCount": 1}, "repositoryOnDiskBytes": 1233101396}, "outcome": "failed", "status": "success", "submitTime": "2026-09-21T09:19:32.876555+00:00", "targetGpu": "Kunlunxin_p-800", "taskId": "4997864", "taskType": "text-generation", "verifyResult": -1}
+{"failReason": "ambiguous_runtime", "failureAction": "send_compact_profile_and_root_exception_to_llm", "failureCategory": "ambiguous_runtime", "failureClassificationReason": "execute_empty_result", "failureCode": "EXECUTE_EMPTY_RESULT", "failureDeterministic": false, "failureEnrichmentAttempts": 1, "failureEnrichmentError": null, "failureNeedsLlm": true, "failureScope": "unknown", "framework": "vllm-mlu", "lastSyncTime": "2026-09-27T13:43:22.766786+00:00", "modelId": "RedHatAI/Mistral-Nemo-Instruct-2407-quantized.w8a16", "modelProfile": {"architectures": ["MistralForCausalLM"], "configFingerprint": "36ec94e289a8da5fd70b98a2f401fc5159740daf9e2a076da4fd490f34bcbc0b", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 13594088584, "estimatedRequiredGiB": 15.203, "gpuMemoryEvidence": {"memoryGiB": 24.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 24.0, "maximumRepositorySizeGiB": 20.0, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "mistral", "modelscopeFileSize": 13603621925, "modelscopeLicense": "apache-2.0", "modelscopeParams": 12247782400, "modelscopeTags": ["license:apache-2.0", "model_type:mistral", "library:transformer", "library:safetensors", "library:pytorch", "task:text-generation"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "compressed-tensors", "repositoryOnDiskBytes": 13603621925}, "outcome": "failed", "status": "success", "submitTime": "2026-09-21T09:19:32.873304+00:00", "targetGpu": "Cambricon_mlu-370-x4", "taskId": "4997863", "taskType": "text-generation", "verifyResult": -1}
 {"failReason": "platform_infrastructure", "failureAction": "open_short_gpu_framework_circuit_and_retry_other_models", "failureCategory": "platform_infrastructure", "failureClassificationReason": "platform_io_transient", "failureCode": "STORAGE_ERROR", "failureDetectedFramework": "vllm", "failureDeterministic": true, "failureEnrichmentAttempts": 1, "failureEnrichmentError": null, "failureNeedsLlm": false, "failureScope": "gpu_framework", "framework": "vllm", "lastSyncTime": "2026-09-23T12:29:42.455438+00:00", "modelId": "RedHatAI/Qwen2-1.5B-Instruct-quantized.w8a8", "modelProfile": {"architectures": ["Qwen2ForCausalLM"], "configFingerprint": "2ebd1d3ee8ed952dd08e24120ad5d7342c9a018f55c478225a242b664721e8d0", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 2245331488, "estimatedRequiredGiB": 2.522, "gpuMemoryEvidence": {"memoryGiB": 32.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 32.0, "maximumRepositorySizeGiB": 26.667, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "qwen2", "modelscopeFileSize": 2256941144, "modelscopeLicense": "apache-2.0", "modelscopeParams": 1777088000, "modelscopeTags": ["license:apache-2.0", "model_type:qwen2", "library:transformer", "library:safetensors", "library:pytorch", "task:text-generation"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "compressed-tensors", "repositoryOnDiskBytes": 2256941144}, "outcome": "failed", "status": "success", "submitTime": "2026-09-21T09:13:58.610056+00:00", "targetGpu": "Ascend_910-b4", "taskId": "4997803", "taskType": "text-generation", "verifyResult": -1}
 {"failReason": "ambiguous_runtime", "failureAction": "send_compact_profile_and_root_exception_to_llm", "failureCategory": "ambiguous_runtime", "failureClassificationReason": "driver_error", "failureCode": "DRIVER_ERROR", "failureDetectedFramework": "vllm", "failureDeterministic": false, "failureEnrichmentAttempts": 1, "failureEnrichmentError": null, "failureNeedsLlm": true, "failureScope": "unknown", "framework": "vllm", "lastSyncTime": "2026-09-24T10:43:18.860153+00:00", "modelId": "nm-testing/Meta-Llama-3.1-8B-Instruct-FP8-hf", "modelProfile": {"architectures": ["LlamaForCausalLM"], "configFingerprint": "2ebd1d3ee8ed952dd08e24120ad5d7342c9a018f55c478225a242b664721e8d0", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 9081287016, "estimatedRequiredGiB": 10.159, "gpuMemoryEvidence": {"memoryGiB": 32.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 32.0, "maximumRepositorySizeGiB": 26.667, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "llama", "modelscopeFileSize": 9090504988, "modelscopeLicense": "llama3.1", "modelscopeParams": 8030261248, "modelscopeTags": ["license:llama3.1", "model_type:llama", "library:pytorch", "library:transformer", "library:safetensors", "task:text-generation", "custom_tag:fp8", "custom_tag:vllm"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "compressed-tensors", "repositoryOnDiskBytes": 9090504988}, "outcome": "failed", "status": "success", "submitTime": "2026-09-21T09:13:51.958882+00:00", "targetGpu": "Ascend_910-b4", "taskId": "4997802", "taskType": "text-generation", "verifyResult": -1}
 {"failReason": "ambiguous_runtime", "failureAction": "send_compact_profile_and_root_exception_to_llm", "failureCategory": "ambiguous_runtime", "failureClassificationReason": "driver_error", "failureCode": "DRIVER_ERROR", "failureDeterministic": false, "failureEnrichmentAttempts": 1, "failureEnrichmentError": null, "failureNeedsLlm": true, "failureScope": "unknown", "framework": "vllm-customized", "lastSyncTime": "2026-09-22T01:48:12.547533+00:00", "modelId": "XHToken/Spark-X2.5-1.7B", "modelProfile": {"architectures": ["Spark2_5ForCausalLM"], "configFingerprint": "1b41437401598cde7836ccc25883635c2738b06a5af924b7497b6e747e655b4b", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 3415340496, "estimatedRequiredGiB": 3.834, "gpuMemoryEvidence": {"memoryGiB": 24.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 24.0, "maximumRepositorySizeGiB": 20.0, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "spark2_5", "modelscopeFileSize": 3430847031, "modelscopeLicense": "apache-2.0", "modelscopeParams": 1707657216, "modelscopeTags": ["license:apache-2.0", "model_type:spark2_5", "library:safetensors", "library:", "task:text-generation", "custom_tag:llm", "custom_tag:sparkx2_5"], "modelscopeTasks": ["text-generation"], "quantizationMethod": null, "repositoryOnDiskBytes": 3430847031}, "outcome": "failed", "status": "success", "submitTime": "2026-09-21T09:10:32.259004+00:00", "targetGpu": "Cambricon_mlu-370-x8", "taskId": "4997718", "taskType": "text-generation", "verifyResult": -1}
@@ -297,4 +298,3 @@
 {"failReason": "runtime_memory", "failureAction": "lower_memory_risk_or_reject_combination", "failureCategory": "runtime_memory", "failureClassificationReason": "structured_device_oom", "failureCode": "DEVICE_OOM", "failureDetectedFramework": "vllm", "failureDeterministic": true, "failureEnrichmentAttempts": 1, "failureEnrichmentError": null, "failureNeedsLlm": false, "failureScope": "model_gpu", "framework": "vllm", "lastSyncTime": "2026-09-21T02:38:57.372787+00:00", "modelId": "nv-community/Llama-3_3-Nemotron-Super-49B-v1_5-NVFP4", "modelProfile": {}, "outcome": "failed", "status": "success", "submitTime": "2026-09-21T02:37:21+00:00", "targetGpu": "MetaX_c-500", "taskId": "4079201", "taskType": "text-generation", "verifyResult": -1}
 {"failReason": "ambiguous_runtime", "failureAction": "send_compact_profile_and_root_exception_to_llm", "failureCategory": "ambiguous_runtime", "failureClassificationReason": "driver_error", "failureCode": "DRIVER_ERROR", "failureDetectedFramework": "vllm", "failureDeterministic": false, "failureEnrichmentAttempts": 1, "failureEnrichmentError": null, "failureNeedsLlm": true, "failureScope": "unknown", "framework": "vllm", "lastSyncTime": "2026-09-22T07:42:35.356592+00:00", "modelId": "solidrust/Llama-3-16B-Instruct-v0.1-AWQ", "modelProfile": {"architectures": ["LlamaForCausalLM"], "configFingerprint": "53aadead5bb69dd6c72976cf6353832080116244f86ffae86297646e5bb48fdc", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 10261168632, "estimatedRequiredGiB": 11.478, "gpuMemoryEvidence": {"memoryGiB": 64.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 64.0, "maximumRepositorySizeGiB": 53.333, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "llama", "modelscopeFileSize": 10270608122, "modelscopeLicense": "other", "modelscopeParams": 16754741248, "modelscopeTags": ["license:other", "model_type:llama", "library:safetensors", "library:pytorch", "task:text-generation", "custom_tag:4-bit", "custom_tag:AWQ", "custom_tag:text-generation", "custom_tag:autotrain_compatible", "custom_tag:endpoints_compatible - mergekit - merge - facebook - meta - pytorch - llama - llama-3"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "awq", "recentProfileFeedback": {"attributableFailureCount": 0, "consecutiveFailures": 0, "decisionFailureRate": 0.0, "decisionSuccessRate": 0.0, "decisionTotal": 0, "failureBreakdown": {"ambiguous_runtime": 5}, "failureCount": 5, "failureRate": 1.0, "framework": "vllm", "lastTerminalAt": "2026-09-20T22:43:47.754849+00:00", "modelType": "llama", "pendingCount": 0, "pendingRate": 0.0, "platformFailureCount": 0, "quantizationMethod": "awq", "successCount": 0, "successRate": 0.0, "targetGpu": "Ascend_910-b3", "taskType": "text-generation", "total": 5, "unresolvedFailureCount": 5}, "repositoryOnDiskBytes": 10270608122}, "outcome": "failed", "status": "success", "submitTime": "2026-09-21T02:21:28.295232+00:00", "targetGpu": "Ascend_910-b3", "taskId": "4992047", "taskType": "text-generation", "verifyResult": -1}
 {"failReason": "ambiguous_runtime", "failureAction": "send_compact_profile_and_root_exception_to_llm", "failureCategory": "ambiguous_runtime", "failureClassificationReason": "execute_empty_result", "failureCode": "EXECUTE_EMPTY_RESULT", "failureDeterministic": false, "failureEnrichmentAttempts": 1, "failureEnrichmentError": null, "failureNeedsLlm": true, "failureScope": "unknown", "framework": "vllm-customized", "lastSyncTime": "2026-09-21T23:49:30.866184+00:00", "modelId": "OpenBMB/MiniCPM5-2B-MLX", "modelProfile": {"architectures": ["LlamaForCausalLM"], "configFingerprint": "1b41437401598cde7836ccc25883635c2738b06a5af924b7497b6e747e655b4b", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 1416035216, "estimatedRequiredGiB": 1.594, "gpuMemoryEvidence": {"memoryGiB": 24.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 24.0, "maximumRepositorySizeGiB": 20.0, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "llama", "modelscopeFileSize": 1426233716, "modelscopeLicense": "apache-2.0", "modelscopeParams": 393390080, "modelscopeTags": ["license:apache-2.0", "model_type:llama", "library:transformer", "library:safetensors", "task:text-generation", "custom_tag:minicpm", "custom_tag:minicpm5", "custom_tag:llama", "custom_tag:text-generation", "custom_tag:long-context", "custom_tag:tool-calling", "custom_tag:on-device", "custom_tag:edge-ai"], "modelscopeTasks": ["text-generation"], "quantizationMethod": null, "repositoryOnDiskBytes": 1426233716}, "outcome": "failed", "status": "success", "submitTime": "2026-09-21T02:10:00.812814+00:00", "targetGpu": "Cambricon_mlu-370-x8", "taskId": "4991872", "taskType": "text-generation", "verifyResult": -1}
-{"failReason": "ambiguous_runtime", "failureAction": "send_compact_profile_and_root_exception_to_llm", "failureCategory": "ambiguous_runtime", "failureClassificationReason": "execute_empty_result", "failureCode": "EXECUTE_EMPTY_RESULT", "failureDeterministic": false, "failureEnrichmentAttempts": 1, "failureEnrichmentError": null, "failureNeedsLlm": true, "failureScope": "unknown", "framework": "vllm", "lastSyncTime": "2026-09-25T09:35:33.057138+00:00", "modelId": "RedHatAI/gemma-2-9b-it-quantized.w4a16", "modelProfile": {"architectures": ["Gemma2ForCausalLM"], "configFingerprint": "ef4c08c913d3f64cae22550f0c4a189795901d4b11e26d8e640e07e857659fb9", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 7963207224, "estimatedRequiredGiB": 8.924, "gpuMemoryEvidence": {"memoryGiB": 24.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 24.0, "maximumRepositorySizeGiB": 20.0, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "gemma2", "modelscopeFileSize": 7985124576, "modelscopeLicense": "llama2", "modelscopeParams": 10159209984, "modelscopeTags": ["license:llama2", "model_type:gemma2", "library:transformer", "library:safetensors", "library:pytorch", "task:text-generation"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "compressed-tensors", "repositoryOnDiskBytes": 7985124576}, "outcome": "failed", "status": "success", "submitTime": "2026-09-21T02:10:00.705684+00:00", "targetGpu": "Cambricon_mlu-370-x4", "taskId": "4991871", "taskType": "text-generation", "verifyResult": -1}
diff --git a/.modelhub_state/recovery_intents.jsonl b/.modelhub_state/recovery_intents.jsonl
index 41cd29a0a4..c96fb57929 100644
--- a/.modelhub_state/recovery_intents.jsonl
+++ b/.modelhub_state/recovery_intents.jsonl
@@ -2549,6 +2549,7 @@
 {"batchId": "b904c4874f404e5eaaa48ab07d59d26f", "completedAt": "2026-09-27T08:40:31.922775+00:00", "configFingerprint": "a50e4651a42a3470debcc065c9742ec9ff939d991aa62aea3b27f6df6fc913ba", "configSource": "modelhub_live", "createdAt": "2026-09-27T08:39:40.171340+00:00", "framework": "llamacpp", "intentId": "ff057c0513364b4fbc832d52f687c3b4", "lastModified": "2026-09-23T16:50:00+00:00", "modelAddress": "https://modelscope.cn/models/IFM/K2-Horizon-0.9B-GGUF", "reason": null, "reconciledAt": "2026-09-27T13:38:47.219527+00:00", "repoId": "IFM/K2-Horizon-0.9B-GGUF", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 4096}, "status": "recovered_active", "targetGpu": "Ascend_910-b4", "taskId": "5117975", "taskType": "text-generation"}
 {"batchId": "540be06560744831aaa4fa211fcd5719", "completedAt": "2026-09-27T08:57:50.007778+00:00", "configFingerprint": "d7268287472242c2dea130d2ea50f93a8978c6b045245aba0d1278a5ffd7f471", "configSource": "modelhub_live", "createdAt": "2026-09-27T08:57:46.580095+00:00", "framework": "llamacpp", "intentId": "a2340027512348eb9c73a8da0cfe1d9b", "lastModified": "2026-09-23T16:50:00+00:00", "modelAddress": "https://modelscope.cn/models/IFM/K2-Horizon-0.9B-GGUF", "reason": null, "reconciledAt": "2026-09-27T13:38:47.218838+00:00", "repoId": "IFM/K2-Horizon-0.9B-GGUF", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 2048}, "status": "recovered_active", "targetGpu": "Mthreads_s4000", "taskId": "5118194", "taskType": "text-generation"}
 {"batchId": "d80c5996172c485ba3db82f0ed34d979", "completedAt": "2026-09-27T10:23:52.110646+00:00", "configFingerprint": "16f74ea0f409861ab0c8169a25d9484cb482e27d6a3ad7c529a70b7cdbfabee3", "configSource": "modelhub_live", "createdAt": "2026-09-27T10:23:50.304966+00:00", "framework": "vllm_fix_tokenizer", "intentId": "c85cf963d3954293b0fb29a9ecfb7b94", "lastModified": "2026-09-27T05:41:43+00:00", "modelAddress": "https://modelscope.cn/models/FINAL-Bench/Darwin-27B-RSI", "reason": null, "reconciledAt": "2026-09-27T13:38:47.219055+00:00", "repoId": "FINAL-Bench/Darwin-27B-RSI", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 4096}, "status": "recovered_active", "targetGpu": "Kunlunxin_p-800", "taskId": "5119577", "taskType": "text-generation"}
+{"batchId": "7dfff4ac7fb049978110c3e3a01b8030", "configFingerprint": "16f74ea0f409861ab0c8169a25d9484cb482e27d6a3ad7c529a70b7cdbfabee3", "configSource": "modelhub_live", "createdAt": "2026-09-27T13:44:58.286855+00:00", "framework": "vllm_fix_tokenizer", "intentId": "00a5545fd4df410a9210b7174cc4cb13", "lastModified": "2026-09-27T13:41:58+00:00", "modelAddress": "https://modelscope.cn/models/Qwen/Qwen3Guard-Stream-0.6B", "repoId": "Qwen/Qwen3Guard-Stream-0.6B", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 4096}, "status": "pending", "targetGpu": "Kunlunxin_p-800", "taskType": "text-generation"}
 {"batchId": "179e9ecfe35348deb44f3c5be7efe247", "completedAt": "2026-09-26T04:54:18.911541+00:00", "configFingerprint": "16f74ea0f409861ab0c8169a25d9484cb482e27d6a3ad7c529a70b7cdbfabee3", "configSource": "modelhub_live", "createdAt": "2026-09-26T04:54:16.299502+00:00", "framework": "vllm_fix_tokenizer", "intentId": "f015fa059f0d4b85abd1ea1620d3a8cd", "repoId": "mlx-community/LFM2.5-1.2B-Instruct-6bit", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 4096}, "status": "uniqueness_rejected", "targetGpu": "Biren_166m", "taskId": null, "taskType": "text-generation"}
 {"batchId": "f9c8240154db42b398dacf34696c5086", "completedAt": "2026-09-24T11:26:48.472659+00:00", "configFingerprint": "1c0a4f69628b2a0ad37a6eee139a4c98afed7d8c93529a9ef42a90756e4e1940", "configSource": "modelhub_live", "createdAt": "2026-09-24T11:26:46.008294+00:00", "framework": "vllm_tokenizer_patch", "intentId": "10a1c592a9ee4f11bef9a65c25835867", "repoId": "mlx-community/LFM2.5-1.2B-Instruct-6bit", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 4096}, "status": "uniqueness_rejected", "targetGpu": "Kunlunxin_p-800", "taskId": null, "taskType": "text-generation"}
 {"batchId": "72ad98bd76b64878af3b85496f9b4f2a", "completedAt": "2026-09-24T03:18:53.744906+00:00", "configFingerprint": "1566bbd6c69427ad32f7f86aeba4e39850a1a7af5c6f4ef3c9ca82f0e882f3c5", "configSource": "modelhub_live", "createdAt": "2026-09-24T03:18:52.675723+00:00", "framework": "vllm-mlu", "intentId": "bc91c43bd85e44bda5e2f46293875f01", "repoId": "mlx-community/LFM2.5-1.2B-Instruct-6bit", "safeConfigVector": {"dtype": "-", "gpuNum": 1}, "status": "uniqueness_rejected", "targetGpu": "Cambricon_mlu-370-x8", "taskId": null, "taskType": "text-generation"}
diff --git a/manifest.json b/manifest.json
index be90bc6301..8f4cc1e666 100644
--- a/manifest.json
+++ b/manifest.json
@@ -2,24 +2,24 @@
   "agentVersion": "2026.09.22.1",
   "checksums": {
     ".modelhub_state/account_capacity.json": "930f9869793c3d1aa071c53f05b7cf82a4e372f002be88361d6d6189e27c12a1",
-    ".modelhub_state/architecture_compatibility_blacklist.json": "0751ae02bf772ab86b7c7f478f671027c3aea7302fadbe076857bd8c9f29a285",
+    ".modelhub_state/architecture_compatibility_blacklist.json": "5070c8855801b32c54ef3fafec23d97ea8c1966d23c7a4b06cebe07ff1ffc681",
     ".modelhub_state/architecture_history_backfill.json": "a92348206605da04f75c9e26e7c818b76f259e0b28983f18b6375ef94802f0ab",
-    ".modelhub_state/market_intelligence.json": "3729a382e63a5f110a4047b92057690fb659a944232064b122c67848064f0947",
-    ".modelhub_state/official_capabilities.json": "6519659b3b1732313f6a3eca3b054a40eae4ad93ee60af0a53fff37307d07ef0",
-    ".modelhub_state/outcome_checkpoint.json": "1f5578a8f7d3a576862df29804a34e8adc87c59aa151e7a737e063fcdf065367",
+    ".modelhub_state/market_intelligence.json": "0c3d1d535e3d3205b00ad643e814e8af6c79bd56e8f049f553f555df70b2acb2",
+    ".modelhub_state/official_capabilities.json": "53d14aecb2a41331e7055cf7ae0807e5f2124afaecd0041a2a5492b1dd2ee013",
+    ".modelhub_state/outcome_checkpoint.json": "803cbb141df92f7314ddc5a4c242c011e013c5dca6c55b8e98ca670e41be4a9c",
     ".modelhub_state/queue_cleanup_latest.json": "694bb6e81572a5a5f3e554e2c6eff62ff80ba911ec7f055a8070daa352e6f6d3",
-    ".modelhub_state/recent_outcomes.jsonl": "8d7a16bc15a60edabfa565d40751c935d84374b5b03fac0f74e0d6db01689118",
+    ".modelhub_state/recent_outcomes.jsonl": "29de8caa53d34c5cd69118aa9b7b19270cfeb5f2ed4f79592568a96e8c2790ab",
     ".modelhub_state/recovery_active_tasks.jsonl": "9009526204f39358eb6d078127485c8e8ffa27dcf08d4650d3eddc8f09a8d3d0",
-    ".modelhub_state/recovery_intents.jsonl": "16973b3d543bd590db778e2b17b51db205bb47229fab9cb2dab95db041ae4f85",
+    ".modelhub_state/recovery_intents.jsonl": "50a55b19b5fde732260d6c437cc2ece689c1585b73aa799a415dac7c646b50d3",
     ".modelhub_state/routing_intelligence.json": "1db5177b676abb095b389aaa8028a5c17a8088175f566c791334114b3f58f9b4",
     ".modelhub_state/submission_exclusions.jsonl": "2cd0731cd4d31bc6d844f4f0a1e510e2fc839ca66f2ce07d2664110b33374799",
     ".modelhub_state/worker_crashes.jsonl": "8f862697b8c8286a6ac8130827f91487a3eb07ecaf4af7ee3de2915d925deb10",
     "ledger/submissions.jsonl": "3b64aad3276de1e5374b8f3768d62e11184e8dfd4f353ea76f12af9488f38272",
-    "outcomes/submissions.jsonl": "e7b52b89fa2e00a74a557a7acbea6780d6b0e0ea737345eb7f420efbf2514e4a"
+    "outcomes/submissions.jsonl": "0adcb340b1452237288c928624bf2b88e34a76b71e145cca08139f8028a12df4"
   },
-  "generation": 17348,
-  "phase": "cycle",
+  "generation": 17349,
+  "phase": "intent",
   "schemaVersion": 1,
-  "updatedAt": "2026-09-27T13:42:21.362546+00:00",
+  "updatedAt": "2026-09-27T13:44:58.424815+00:00",
   "writerId": "6e8ddfbd92ca45ccaf79298516f8a6ea"
 }
diff --git a/outcomes/submissions.jsonl b/outcomes/submissions.jsonl
index 0ff75d4770..db267e1a00 100644
--- a/outcomes/submissions.jsonl
+++ b/outcomes/submissions.jsonl
@@ -460,7 +460,6 @@
 {"failReason": null, "framework": "vllm_fix_tokenizer", "lastSyncTime": "2026-09-21T17:22:47.256894+00:00", "modelId": "blue2star/Qwen-Image-2.1-PE-I2I-ComfyUI", "modelProfile": {"architectures": ["Qwen3_5ForConditionalGeneration"], "configFingerprint": "16f74ea0f409861ab0c8169a25d9484cb482e27d6a3ad7c529a70b7cdbfabee3", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 29306922460, "estimatedRequiredGiB": 32.775, "gpuMemoryEvidence": {"memoryGiB": 64.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 64.0, "maximumRepositorySizeGiB": 53.333, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "qwen3_5", "modelscopeFileSize": 29326959151, "modelscopeLicense": "other", "modelscopeParams": 18821595084, "modelscopeTags": ["license:other", "model_type:qwen3_5", "library:pytorch", "library:transformer", "library:safetensors", "task:text-generation", "custom_tag:qwen", "custom_tag:prompt-rewriting", "custom_tag:image-editing"], "modelscopeTasks": ["text-generation"], "quantizationMethod": null, "repositoryOnDiskBytes": 29326959151}, "outcome": "pending", "status": "waiting", "submitTime": "2026-09-21T08:59:28.154388+00:00", "targetGpu": "Biren_166m", "taskId": "4997528", "taskType": "text-generation", "verifyResult": null}
 {"failReason": null, "framework": "vllm_tokenizer_patch", "lastSyncTime": "2026-09-21T17:22:47.256900+00:00", "modelId": "nm-testing/Meta-Llama-3-8B-Instruct-W8A8-FP8-Channelwise-compressed-tensors", "modelProfile": {"architectures": ["LlamaForCausalLM"], "configFingerprint": "7b5500238d52e9e2a21471d722bc1636ca047e1a12ba49c491593b86bec602b8", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 9084014480, "estimatedRequiredGiB": 10.162, "gpuMemoryEvidence": {"memoryGiB": 64.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 64.0, "maximumRepositorySizeGiB": 53.333, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "llama", "modelscopeFileSize": 9093198689, "modelscopeLicense": null, "modelscopeParams": 8030261248, "modelscopeTags": ["model_type:llama", "library:safetensors", "library:pytorch", "task:text-generation"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "compressed-tensors", "repositoryOnDiskBytes": 9093198689}, "outcome": "pending", "status": "waiting", "submitTime": "2026-09-21T09:02:30.537080+00:00", "targetGpu": "Ascend_910-b3", "taskId": "4997592", "taskType": "text-generation", "verifyResult": null}
 {"failReason": null, "framework": "vllm-patch-tokenizer", "lastSyncTime": "2026-09-21T17:22:47.256807+00:00", "modelId": "nm-testing/tinyllama-oneshot-w8w8-test-static-shape-change", "modelProfile": {"architectures": ["LlamaForCausalLM"], "configFingerprint": "4d80215b8854e486ecf078eda8f5db649838852684b9429ab6d7b0d91bb411cf", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 1231270112, "estimatedRequiredGiB": 1.378, "gpuMemoryEvidence": {"memoryGiB": 64.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 64.0, "maximumRepositorySizeGiB": 53.333, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "llama", "modelscopeFileSize": 1233120748, "modelscopeLicense": null, "modelscopeParams": 1100048384, "modelscopeTags": ["model_type:llama", "library:safetensors", "library:pytorch", "task:text-generation"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "compressed-tensors", "repositoryOnDiskBytes": 1233120748}, "outcome": "pending", "status": "waiting", "submitTime": "2026-09-21T09:07:46.198212+00:00", "targetGpu": "hygon_k100-ai", "taskId": "4997684", "taskType": "text-generation", "verifyResult": null}
-{"failReason": null, "framework": "vllm-mlu", "lastSyncTime": "2026-09-21T17:22:47.256846+00:00", "modelId": "RedHatAI/Mistral-Nemo-Instruct-2407-quantized.w8a16", "modelProfile": {"architectures": ["MistralForCausalLM"], "configFingerprint": "36ec94e289a8da5fd70b98a2f401fc5159740daf9e2a076da4fd490f34bcbc0b", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 13594088584, "estimatedRequiredGiB": 15.203, "gpuMemoryEvidence": {"memoryGiB": 24.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 24.0, "maximumRepositorySizeGiB": 20.0, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "mistral", "modelscopeFileSize": 13603621925, "modelscopeLicense": "apache-2.0", "modelscopeParams": 12247782400, "modelscopeTags": ["license:apache-2.0", "model_type:mistral", "library:transformer", "library:safetensors", "library:pytorch", "task:text-generation"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "compressed-tensors", "repositoryOnDiskBytes": 13603621925}, "outcome": "pending", "status": "waiting", "submitTime": "2026-09-21T09:19:32.873304+00:00", "targetGpu": "Cambricon_mlu-370-x4", "taskId": "4997863", "taskType": "text-generation", "verifyResult": null}
 {"failReason": null, "framework": "vllm", "lastSyncTime": "2026-09-21T17:22:47.256927+00:00", "modelId": "nm-testing/tinyllama-oneshot-w8a8-static-v2", "modelProfile": {"architectures": ["LlamaForCausalLM"], "configFingerprint": "c25be7d7c1bf173abe564916f8a370d9b648beb3aea2a79522523aaa6b69ee29", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 1231270112, "estimatedRequiredGiB": 1.378, "gpuMemoryEvidence": {"memoryGiB": 64.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 64.0, "maximumRepositorySizeGiB": 53.333, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "llama", "modelscopeFileSize": 1233118759, "modelscopeLicense": null, "modelscopeParams": 1100048384, "modelscopeTags": ["model_type:llama", "library:safetensors", "library:pytorch", "task:text-generation"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "compressed-tensors", "repositoryOnDiskBytes": 1233118759}, "outcome": "pending", "status": "waiting", "submitTime": "2026-09-21T09:19:32.868721+00:00", "targetGpu": "Ascend_910-b3", "taskId": "4997865", "taskType": "text-generation", "verifyResult": null}
 {"failReason": null, "framework": "vllm_tokenizer_patch", "lastSyncTime": "2026-09-21T17:41:26.757131+00:00", "modelId": "nm-testing/Meta-Llama-3-8B-Instruct-W8A8-Dyn-Per-Token", "modelProfile": {"architectures": ["LlamaForCausalLM"], "configFingerprint": "7b5500238d52e9e2a21471d722bc1636ca047e1a12ba49c491593b86bec602b8", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 9084013064, "estimatedRequiredGiB": 10.162, "gpuMemoryEvidence": {"memoryGiB": 64.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 64.0, "maximumRepositorySizeGiB": 53.333, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "llama", "modelscopeFileSize": 9093197403, "modelscopeLicense": null, "modelscopeParams": 8030261248, "modelscopeTags": ["model_type:llama", "library:safetensors", "library:pytorch", "task:text-generation"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "compressed-tensors", "repositoryOnDiskBytes": 9093197403}, "outcome": "pending", "status": "waiting", "submitTime": "2026-09-21T09:40:34.490655+00:00", "targetGpu": "Ascend_910-b3", "taskId": "4998300", "taskType": "text-generation", "verifyResult": null}
 {"failReason": null, "framework": "vllm_tokenizer_patch", "lastSyncTime": "2026-09-21T18:00:38.063615+00:00", "modelId": "RedHatAI/gemma-2-2b-quantized.w8a16", "modelProfile": {"architectures": ["Gemma2ForCausalLM"], "configFingerprint": "7b5500238d52e9e2a21471d722bc1636ca047e1a12ba49c491593b86bec602b8", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 6746735136, "estimatedRequiredGiB": 7.565, "gpuMemoryEvidence": {"memoryGiB": 64.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 64.0, "maximumRepositorySizeGiB": 53.333, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "gemma2", "modelscopeFileSize": 6768615358, "modelscopeLicense": "gemma", "modelscopeParams": 3204165888, "modelscopeTags": ["license:gemma", "model_type:gemma2", "library:transformer", "library:safetensors", "library:pytorch", "task:text-generation", "custom_tag:int8", "custom_tag:vllm"], "modelscopeTasks": ["text-generation"], "quantizationMethod": "compressed-tensors", "repositoryOnDiskBytes": 6768615358}, "outcome": "pending", "status": "waiting", "submitTime": "2026-09-21T09:41:35.705044+00:00", "targetGpu": "Ascend_910-b3", "taskId": "4998344", "taskType": "text-generation", "verifyResult": null}