diff --git a/.modelhub_state/recovery_intents.jsonl b/.modelhub_state/recovery_intents.jsonl
index cd0c236687..25bc672e15 100644
--- a/.modelhub_state/recovery_intents.jsonl
+++ b/.modelhub_state/recovery_intents.jsonl
@@ -2692,7 +2692,7 @@
 {"batchId": "b4bfc4da77d345d8a99cac1b8c43f567", "completedAt": "2026-10-03T18:55:01.132632+00:00", "configFingerprint": "16f74ea0f409861ab0c8169a25d9484cb482e27d6a3ad7c529a70b7cdbfabee3", "configSource": "modelhub_live", "createdAt": "2026-10-03T18:53:07.104151+00:00", "framework": "vllm_fix_tokenizer", "intentId": "faf445fca7f44c759a23a50eabcbf299", "lastModified": "2026-10-03T18:52:06+00:00", "modelAddress": "https://modelscope.cn/models/LessThanThreeAI/Qwen3.8-27B-Humanlike-Chat-2.0-GPTQ-Int4", "reason": null, "reconciledAt": "2026-10-04T18:36:10.569690+00:00", "repoId": "LessThanThreeAI/Qwen3.8-27B-Humanlike-Chat-2.0-GPTQ-Int4", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 4096}, "status": "recovered_active", "targetGpu": "Biren_166m", "taskId": "5263453", "taskType": "text-generation"}
 {"batchId": "f080e6c52a01419faf86ef7c25efb6b0", "completedAt": "2026-10-03T19:12:17.015228+00:00", "configFingerprint": "36ec94e289a8da5fd70b98a2f401fc5159740daf9e2a076da4fd490f34bcbc0b", "configSource": "modelhub_live", "createdAt": "2026-10-03T19:12:14.079992+00:00", "framework": "vllm-mlu", "intentId": "73a271af9f3d490182efdd017192cab2", "lastModified": "2026-10-03T18:52:06+00:00", "modelAddress": "https://modelscope.cn/models/LessThanThreeAI/Qwen3.8-27B-Humanlike-Chat-2.0-GPTQ-Int4", "reason": null, "reconciledAt": "2026-10-04T18:36:10.569687+00:00", "repoId": "LessThanThreeAI/Qwen3.8-27B-Humanlike-Chat-2.0-GPTQ-Int4", "safeConfigVector": {"dtype": "-", "gpuNum": 1}, "status": "recovered_active", "targetGpu": "Cambricon_mlu-370-x4", "taskId": "5264047", "taskType": "text-generation"}
 {"batchId": "3cefb7c4f2d34806a07a139e53bbd256", "completedAt": "2026-10-04T01:54:08.997071+00:00", "configFingerprint": "16f74ea0f409861ab0c8169a25d9484cb482e27d6a3ad7c529a70b7cdbfabee3", "configSource": "modelhub_live", "createdAt": "2026-10-04T01:52:21.091751+00:00", "framework": "vllm_fix_tokenizer", "intentId": "5d86c1ebc5364fa5806a6b876e7cbd8b", "lastModified": "2026-10-03T15:25:47+00:00", "modelAddress": "https://modelscope.cn/models/SC117/Ornith-1.0-35B-Heretic-MTP", "reason": null, "reconciledAt": "2026-10-04T18:36:10.569685+00:00", "repoId": "SC117/Ornith-1.0-35B-Heretic-MTP", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 4096}, "status": "recovered_active", "targetGpu": "Kunlunxin_p-800", "taskId": "5270463", "taskType": "text-generation"}
-{"batchId": "2adc2e90477d48d7a40c07fb98ab6fb3", "configFingerprint": "5f9f3c1607c4d592480a2bc6bf4bd3811708695139f6772f4b18a577c41a71ec", "configSource": "modelhub_live", "createdAt": "2026-10-04T18:39:57.296004+00:00", "framework": "vllm_tokenizer_patch", "intentId": "02632a5d05a648f398218c1ea61d76c3", "lastModified": "2026-10-04T18:30:07+00:00", "modelAddress": "https://modelscope.cn/models/bench-labs/cagliostro-v3.5", "repoId": "bench-labs/cagliostro-v3.5", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 2048}, "status": "pending", "targetGpu": "MetaX_c-500", "taskType": "text-generation"}
+{"batchId": "2adc2e90477d48d7a40c07fb98ab6fb3", "completedAt": "2026-10-04T18:41:35.313675+00:00", "configFingerprint": "5f9f3c1607c4d592480a2bc6bf4bd3811708695139f6772f4b18a577c41a71ec", "configSource": "modelhub_live", "createdAt": "2026-10-04T18:39:57.296004+00:00", "framework": "vllm_tokenizer_patch", "intentId": "02632a5d05a648f398218c1ea61d76c3", "lastModified": "2026-10-04T18:30:07+00:00", "modelAddress": "https://modelscope.cn/models/bench-labs/cagliostro-v3.5", "reason": null, "repoId": "bench-labs/cagliostro-v3.5", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 2048}, "status": "submitted", "targetGpu": "MetaX_c-500", "taskId": "5285150", "taskType": "text-generation"}
 {"batchId": "9544bf0548e54a8e93c3adfd9bffb52a", "completedAt": "2026-10-03T16:15:51.005765+00:00", "configFingerprint": "0d5bc82787a89d4313672808d1a8629a1fc675c91aa6a08865f770aa24447fbc", "configSource": "modelhub_live", "createdAt": "2026-10-03T16:13:49.532883+00:00", "framework": "vllm-patch-tokenizer", "intentId": "7f7cb1b51f944ac58018ee2e42ca61f7", "repoId": "mlx-community/LFM2.5-1.2B-Instruct-6bit", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 4096}, "status": "uniqueness_rejected", "targetGpu": "hygon_k100-ai", "taskId": null, "taskType": "text-generation"}
 {"batchId": "6ad67bedd5104b958247ed686fa4a8d7", "completedAt": "2026-09-29T14:16:48.925444+00:00", "configFingerprint": "8be5b78bdba5c4413c1e4ab91a46e27c34642b25d5dc059ad3a11d50d7e33593", "configSource": "modelhub_live", "createdAt": "2026-09-29T14:16:45.683768+00:00", "framework": "vllm_tokenizer_patch", "intentId": "ba30792e89b44ef5804d5f3613d7df71", "repoId": "nv-community/Qwen3.8-27B-NVFP4", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 4096}, "status": "uniqueness_rejected", "targetGpu": "MetaX_c-500", "taskId": null, "taskType": "text-generation"}
 {"batchId": "6ad67bedd5104b958247ed686fa4a8d7", "completedAt": "2026-09-29T14:16:48.925408+00:00", "configFingerprint": "8be5b78bdba5c4413c1e4ab91a46e27c34642b25d5dc059ad3a11d50d7e33593", "configSource": "modelhub_live", "createdAt": "2026-09-29T14:16:45.683638+00:00", "framework": "vllm_tokenizer_patch", "intentId": "69aec4d25f21464c860135eb90d77735", "repoId": "nv-community/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8", "safeConfigVector": {"gpuNum": 1, "maxModelLen": 4096}, "status": "uniqueness_rejected", "targetGpu": "MetaX_c-500", "taskId": null, "taskType": "text-generation"}
diff --git a/.modelhub_state/routing_intelligence.json b/.modelhub_state/routing_intelligence.json
index b63a12229b..e14865a1f5 100644
--- a/.modelhub_state/routing_intelligence.json
+++ b/.modelhub_state/routing_intelligence.json
@@ -1,6 +1,6 @@
 {
   "acceptedByCategory": {
-    "unified_success_first": 2831
+    "unified_success_first": 2832
   },
   "acceptedByRoute": {
     "text-generation|Ascend_910-b3|llamacpp": 26,
@@ -29,7 +29,7 @@
     "text-generation|Kunlunxin_p-800|vllm_fix_tokenizer": 239,
     "text-generation|Kunlunxin_p-800|vllm_tokenizer_patch": 98,
     "text-generation|MetaX_c-500|vllm": 127,
-    "text-generation|MetaX_c-500|vllm_tokenizer_patch": 17,
+    "text-generation|MetaX_c-500|vllm_tokenizer_patch": 18,
     "text-generation|Mthreads_s4000|llamacpp": 21,
     "text-generation|Mthreads_s4000|vllm": 123,
     "text-generation|Sunrise_pt-200-x1|vllm": 87,
@@ -39,8 +39,8 @@
     "text-generation|hygon_k100-ai|vllm": 27,
     "text-generation|hygon_k100-ai|vllm-patch-tokenizer": 153
   },
-  "acceptedSinceRefresh": 2831,
-  "acceptedTotal": 2831,
-  "generatedAt": "2026-10-04T01:54:08.953377+00:00",
+  "acceptedSinceRefresh": 2832,
+  "acceptedTotal": 2832,
+  "generatedAt": "2026-10-04T18:41:35.269842+00:00",
   "version": 1
 }
diff --git a/ledger/submissions.jsonl b/ledger/submissions.jsonl
index 517b96b921..50a0820af3 100644
--- a/ledger/submissions.jsonl
+++ b/ledger/submissions.jsonl
@@ -956,3 +956,4 @@
 {"framework": "transformers", "modelAddress": "https://modelscope.cn/models/RedHatAI/starcoder2-15b-quantized.w8a16", "modelId": "RedHatAI/starcoder2-15b-quantized.w8a16", "submitTime": "2026-09-21T18:32:49.780633+00:00", "targetGpu": "Iluvatar_bi-150", "taskId": "5005549", "taskType": "text-generation", "templateId": "modelhub-live-text-generation-transformers-iluvatar-bi-150"}
 {"framework": "vllm_tokenizer_patch", "modelAddress": "https://modelscope.cn/models/mlx-community/gpt-oss-20b-OptiQ-4bit", "modelId": "mlx-community/gpt-oss-20b-OptiQ-4bit", "submitTime": "2026-09-26T08:01:50.847636+00:00", "targetGpu": "Ascend_910-b4", "taskId": "5097139", "taskType": "text-generation", "templateId": "modelhub-live-text-generation-vllm-tokenizer-patch-ascend-910-b4"}
 {"framework": "vllm_fix_tokenizer", "modelAddress": "https://modelscope.cn/models/mlx-community/gpt-oss-20b-OptiQ-4bit", "modelId": "mlx-community/gpt-oss-20b-OptiQ-4bit", "submitTime": "2026-09-26T08:18:32.495861+00:00", "targetGpu": "Kunlunxin_p-800", "taskId": "5097281", "taskType": "text-generation", "templateId": "modelhub-live-text-generation-vllm-fix-tokenizer-kunlunxin-p-800"}
+{"framework": "vllm_tokenizer_patch", "modelAddress": "https://modelscope.cn/models/bench-labs/cagliostro-v3.5", "modelId": "bench-labs/cagliostro-v3.5", "submitTime": "2026-10-04T18:41:34.446649+00:00", "targetGpu": "MetaX_c-500", "taskId": "5285150", "taskType": "text-generation", "templateId": "modelhub-live-text-generation-vllm-tokenizer-patch-metax-c-500"}
diff --git a/manifest.json b/manifest.json
index 22c2294030..e240784f9d 100644
--- a/manifest.json
+++ b/manifest.json
@@ -10,16 +10,16 @@
     ".modelhub_state/queue_cleanup_latest.json": "ec46a79460084d2e9c115ffcb729d6af492c411abe9a039d33ed73d895f9a9f6",
     ".modelhub_state/recent_outcomes.jsonl": "74684e0cb636f08121ef1ea1446e52ebb2f1891266683fdcded66873c8d76476",
     ".modelhub_state/recovery_active_tasks.jsonl": "82d396c79216b7b2ca6815fffdb8a4fec5927172c483b68510c1cc805a1de645",
-    ".modelhub_state/recovery_intents.jsonl": "aae2851a050ca8fef07bd43de5fc1fea56b25e202f4f20ec7112d09bf7d36f72",
-    ".modelhub_state/routing_intelligence.json": "46d123c1283f8ddc7b760ee11a007b1cb9cd850eec877013d71700b3f023d788",
+    ".modelhub_state/recovery_intents.jsonl": "3c0f31494e04b104c35ba2981cae19a000073d12bdb01c5760c07378df505f65",
+    ".modelhub_state/routing_intelligence.json": "d67e87e37cbf08a9d14cff96bde618817ba83008657f92fc0fba5fa8bc844222",
     ".modelhub_state/submission_exclusions.jsonl": "8a533753b09639322718ffe59fad6a6f11e8cc3741d7be12085a149c97ca5507",
     ".modelhub_state/worker_crashes.jsonl": "50ffe2fb7a8d9d9967ff97b47af26f97753a0bbeae8a68113a221474751d9c2d",
-    "ledger/submissions.jsonl": "d8a513a802614cc129a79fd1cbbe72cc8928ceec43ca6017be689b19a789ee75",
-    "outcomes/submissions.jsonl": "e136794593371bae97a38a7e4b4bf6fec9f12c7ecd992bfe69c8d91d5494bd1f"
+    "ledger/submissions.jsonl": "5aa7fd884d9dba4fc9503d22c2aa5f909ce1091622fbc69eb425e344e355beaa",
+    "outcomes/submissions.jsonl": "34161f2524666fe3b668a62c14201d97abcb3a28b091b00ea04806d8b7b1a217"
   },
-  "generation": 20531,
-  "phase": "intent",
+  "generation": 20532,
+  "phase": "result",
   "schemaVersion": 1,
-  "updatedAt": "2026-10-04T18:39:57.485042+00:00",
+  "updatedAt": "2026-10-04T18:41:35.416095+00:00",
   "writerId": "2a476b5b64aa45219c00032efc0102bb"
 }
diff --git a/outcomes/submissions.jsonl b/outcomes/submissions.jsonl
index 3dd1ab008f..52bb160b16 100644
--- a/outcomes/submissions.jsonl
+++ b/outcomes/submissions.jsonl
@@ -748,3 +748,4 @@
 {"failReason": null, "framework": "vllm_fix_tokenizer", "lastSyncTime": "2026-10-04T03:01:41.555967+00:00", "modelId": "LessThanThreeAI/Qwen3.8-27B-Humanlike-Chat-2.0-GPTQ-Int4", "modelProfile": {"architectures": ["Qwen3_5ForConditionalGeneration"], "configFingerprint": "16f74ea0f409861ab0c8169a25d9484cb482e27d6a3ad7c529a70b7cdbfabee3", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 20593110848, "estimatedRequiredGiB": 23.04, "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": null, "modelscopeLicense": "apache-2.0", "modelscopeParams": 424699392, "modelscopeTags": ["license:apache-2.0", "model_type:qwen3_5", "library:safetensors", "library:pytorch", "task:text-generation", "custom_tag:vllm", "custom_tag:sglang", "custom_tag:gptq", "custom_tag:int4", "custom_tag:w4a16", "custom_tag:compressed-tensors", "custom_tag:llm-compressor", "custom_tag:safetensors", "custom_tag:qwen3.8", "custom_tag:conversational", "custom_tag:tool-calling", "custom_tag:function-calling", "custom_tag:roleplay", "custom_tag:character", "custom_tag:humanlike", "custom_tag:uncensored"], "modelscopeTasks": ["text-generation"], "publishedSizeBytes": null, "quantizationMethod": "compressed-tensors", "recursiveRepositoryBytes": 20616225866, "repositoryOnDiskBytes": 20616225866}, "outcome": "pending", "status": "waiting", "submitTime": "2026-10-03T18:55:00.153550+00:00", "targetGpu": "Biren_166m", "taskId": "5263453", "taskType": "text-generation", "verifyResult": null}
 {"failReason": null, "framework": "vllm-mlu", "lastSyncTime": "2026-10-04T03:19:12.246914+00:00", "modelId": "LessThanThreeAI/Qwen3.8-27B-Humanlike-Chat-2.0-GPTQ-Int4", "modelProfile": {"architectures": ["Qwen3_5ForConditionalGeneration"], "configFingerprint": "36ec94e289a8da5fd70b98a2f401fc5159740daf9e2a076da4fd490f34bcbc0b", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 20593110848, "estimatedRequiredGiB": 23.04, "gpuMemoryEvidence": {"memoryGiB": 24.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 24.0, "maximumRepositorySizeGiB": 20.0, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "qwen3_5", "modelscopeFileSize": null, "modelscopeLicense": "apache-2.0", "modelscopeParams": 424699392, "modelscopeTags": ["license:apache-2.0", "model_type:qwen3_5", "library:safetensors", "library:pytorch", "task:text-generation", "custom_tag:vllm", "custom_tag:sglang", "custom_tag:gptq", "custom_tag:int4", "custom_tag:w4a16", "custom_tag:compressed-tensors", "custom_tag:llm-compressor", "custom_tag:safetensors", "custom_tag:qwen3.8", "custom_tag:conversational", "custom_tag:tool-calling", "custom_tag:function-calling", "custom_tag:roleplay", "custom_tag:character", "custom_tag:humanlike", "custom_tag:uncensored"], "modelscopeTasks": ["text-generation"], "publishedSizeBytes": null, "quantizationMethod": "compressed-tensors", "recursiveRepositoryBytes": 20616225866, "repositoryOnDiskBytes": 20616225866}, "outcome": "pending", "status": "waiting", "submitTime": "2026-10-03T19:12:16.744930+00:00", "targetGpu": "Cambricon_mlu-370-x4", "taskId": "5264047", "taskType": "text-generation", "verifyResult": null}
 {"failReason": null, "framework": "vllm_fix_tokenizer", "lastSyncTime": "2026-10-04T10:01:33.651736+00:00", "modelId": "SC117/Ornith-1.0-35B-Heretic-MTP", "modelProfile": {"architectures": ["Qwen3_5MoeForConditionalGeneration"], "configFingerprint": "16f74ea0f409861ab0c8169a25d9484cb482e27d6a3ad7c529a70b7cdbfabee3", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 71908134368, "estimatedRequiredGiB": 80.398, "gpuMemoryEvidence": {"memoryGiB": 100.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 100.0, "maximumRepositorySizeGiB": 83.333, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "qwen3_5_moe", "modelscopeFileSize": 71938541955, "modelscopeLicense": "mit", "modelscopeParams": 35951822704, "modelscopeTags": ["license:mit", "model_type:qwen3_5_moe", "library:safetensors", "library:pytorch", "task:text-generation", "custom_tag:qwen3_5_moe", "custom_tag:reasoning", "custom_tag:agentic-coding", "custom_tag:heretic", "custom_tag:mtp", "custom_tag:apex", "custom_tag:quantization", "custom_tag:gguf", "custom_tag:multimodal", "deploy:swingdeploy"], "modelscopeTasks": ["text-generation"], "publishedSizeBytes": 71938541955, "quantizationMethod": null, "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": "2026-09-23T21:42:41.557218+00:00", "modelType": "qwen3_5_moe", "pendingCount": 0, "pendingRate": 0.0, "platformFailureCount": 0, "quantizationMethod": "none", "successCount": 0, "successRate": 0.0, "targetGpu": "Kunlunxin_p-800", "taskType": "text-generation", "total": 1, "unresolvedFailureCount": 1}, "recursiveRepositoryBytes": 71938541955, "repositoryOnDiskBytes": 71938541955}, "outcome": "pending", "status": "waiting", "submitTime": "2026-10-04T01:54:06.844290+00:00", "targetGpu": "Kunlunxin_p-800", "taskId": "5270463", "taskType": "text-generation", "verifyResult": null}
+{"failReason": null, "framework": "vllm_tokenizer_patch", "lastSyncTime": null, "modelId": "bench-labs/cagliostro-v3.5", "modelProfile": {"architectures": ["CagliostroForCausalLM"], "configFingerprint": "5f9f3c1607c4d592480a2bc6bf4bd3811708695139f6772f4b18a577c41a71ec", "configOptimization": {"applied": false, "source": "official"}, "configSource": "modelhub_live", "estimatedLoadBytes": 585441264, "estimatedRequiredGiB": 0.677, "gpuMemoryEvidence": {"memoryGiB": 64.0, "source": "local_modelhub_preflight_oom"}, "gpuMemoryGiB": 64.0, "maximumRepositorySizeGiB": 53.333, "memorySizingBasis": "recursive_repository_on_disk", "modelCard": {}, "modelType": "cagliostro", "modelscopeFileSize": null, "modelscopeLicense": "apache-2.0", "modelscopeParams": null, "modelscopeTags": ["license:apache-2.0", "model_type:cagliostro", "library:pytorch", "library:transformer", "library:safetensors", "task:text-generation", "custom_tag:causal-lm", "custom_tag:language-model", "custom_tag:base-model", "custom_tag:small-language-model"], "modelscopeTasks": ["text-generation"], "publishedSizeBytes": null, "quantizationMethod": null, "recursiveRepositoryBytes": 605680752, "repositoryOnDiskBytes": 605680752}, "outcome": "pending", "status": "pending", "submitTime": "2026-10-04T18:41:34.446649+00:00", "targetGpu": "MetaX_c-500", "taskId": "5285150", "taskType": "text-generation", "verifyResult": null}